University of Southern California Dissertations and Theses

Like father, like son? A succession-based explanation for conflict initiation by authoritarian regimes

(USC Thesis Other)

Like father, like son? A succession-based explanation for conflict initiation by authoritarian regimes

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content LIKE FATHER, LIKE SON? A SUCCESSION-BASED EXPLANATION FOR CONFLICT INITIATION BY AUTHORITARIAN REGIMES JIHYUN SHIN DISSERTATION PREPARED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY FACULTY OF THE USC GRADUATE SCHOOL Political Science and International Relations (POIR) University of Southern California May 2019 Committee Members (Chair) David C. Kang, Political Science and International relations (POIR) Brian Rathbun, Political Science and International relations (POIR) Nicholas Weller, Political Science at UCR David Walsh, Psychology Abstract It is a widely held belief that unconstrained personalist leaders tend to be most belligerent in their foreign policy pursuit due to the domestic institutional failure of punishing leaders who dominate the military and the state bureaucratic apparatus. This dissertation shows that not the domestic institutional constraints on the leader, but a leader’s predisposition towards use of force better explains the initiation of international military conflicts by autocrats. Using a panel data from 1946-2010, I argue that hereditary personalist autocrats tend to be less conflict prone than other personalist leaders even under similar institutional constraints. The “Power Suc- cessors” whose power acquisition came through hereditary ascription are signif- icantly less conflict-prone than the “Power Seizers” who came to power through use of force even under similar institutional conditions. Furthermore, this disser- tation employs two case studies - Taiwan as a typical case and North Korea as a divergent case - and demonstrates that the divergent paths to power experienced by the “Power Seizers” and the “Power Successors” explain and shape leaders’ beliefs about use of force. 1 Contents Chapter 1: Introduction 7 Chapter 2: Literature Review: Domestic Institutional Constraints and Inter- national Conflict Initiation by Authoritarian Leaders 19 Chapter 3: A Theoretical Framework on Leadership Type and Conflict Initia- tion: A Succession-Based Explanation 26 Three Main Hypotheses: Power Seizers vs. Power Successors vs. Democratic Leaders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Hypothesis: Power Seizers vs. Power Successors . . . . . . . . . . . . 29 Hypothesis: Power Seizers vs. Democratic Leaders . . . . . . . . . . . 31 Hypothesis: Power Successors vs. Democratic Leaders . . . . . . . . . 31 Additional Hypotheses Testing: Power Seizers vs. Leaders under Similar Constraints, Power Successors vs. Leaders under Similar Constraints . . 32 Hypothesis: Power Seizers vs. Autocrats Facing Similar Constraints . . 32 Hypothesis: Power Successors vs. Autocrats Facing Similar Constraints 33 Hypothesis: Hereditary Power Seizers vs. Nonconstitutional Power Successors . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Chapter 4: Measurement and Operationalization of Main IVs and DV 35 Measuring Domestic Institutional Constraints: Are the Power Seizers and the Power Successors Equally Personalist Leaders? . . . . . . . . . . . . . 35 Operationalizing Main IVs: Power Seizers and Power Successors . . . . . . . 42 Power Seizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Constitutional Power Successors . . . . . . . . . . . . . . . . . . . . . 43 Nonconstitutional Power Successors . . . . . . . . . . . . . . . . . . . 45 Dependent Variable and Model Specification . . . . . . . . . . . . . . . . . . 46 Robustness Check . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Alternative Explanation: Domestic Stability . . . . . . . . . . . . . . . 46 Reverse Causation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Standard Control Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Major Power Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Alliances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Contiguity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Openness to Trade . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Regime Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Region Dummies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 2 Chapter 5: Empirical Results 51 H1: Power Seizers vs. Power Successors: The Power Seizers are more likely to initiate international conflicts than the Power Successors. . . . . . . 51 H3a: "Nonconstitutional" Power Successors vs. "Constitutional" Power Suc- cessors: The ”nonconstitutional” Power Successors are more likely to initiate international conflicts than the ”constitutional” Power Successors. 54 H2: Power Seizers vs. Democratic Leaders: The Power Seizers are more likely to initiate international conflicts than democratic leaders. . . . . . 55 H3: Power Successors vs. Democratic Leaders: The Power Successors are no more likely to initiate international conflicts than democratic leaders. . . 56 H4: Power Seizers vs. All Other Autocrats Under Similar Constraints: The Power Seizers are more likely to initiate international conflicts than all other autocrats facing similar institutional constraints. . . . . . . . . . . 58 H5: Power Successors vs. All Other Autocrats Under Similar Constraints: The Power Successors are less likely to initiate international conflicts than all other autocrats facing similar institutional constraints. . . . . . 62 H6:Distinguishing Power Seizers: The Power Successors in nonconstitutional dynasties are less likely to initiate international conflicts than their im- mediate preceding Power Seizers. . . . . . . . . . . . . . . . . . . . . 65 Chapter 6. Case Selection Strategy: A Typical Case vs. A Deviant Case 69 Case Selection: a Typical Case . . . . . . . . . . . . . . . . . . . . . . . . . 69 Case Selection: a Deviant Case . . . . . . . . . . . . . . . . . . . . . . . . . 74 Chapter 7: Taiwan: A Typical Case 78 Mechanism 1: Divergent Paths to Power . . . . . . . . . . . . . . . . . . . . 79 Chiang Kai-shek: An Undiscouraged Old Soldier . . . . . . . . . . . . 80 Chiang Ching-kuo: A Competent Administrator and Scholar Brings Re- form to Taiwan . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Mechanism 2: Like Father, Like Son? Personality Traits and Foreign Policy . 91 Chiang Kai-shek: Provocative, Stubborn, and Narcissistic, Climbs to the Top Leadership . . . . . . . . . . . . . . . . . . . . . . . . 92 Chiang Ching-kuo: A Unpretentious, Practical Leader Changes Course of Taiwan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Chapter 8: North Korea: A Deviant Case 98 Mechanism 1: Self Selection into Power? . . . . . . . . . . . . . . . . . . . 99 Kim Il-Sung: An Anti-Japanese Guerrilla, Who Rose to Power Through War, Purges and Personality Cult . . . . . . . . . . . . . . . . 101 Kim Jong-Il: Spoiled But Ambitious Successor Works His Way Through Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Mechanism 2: Shared Narcissism and Self-Importance: Dispositional Traits and Foreign Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 3 Kim Il-sung: Fatherless and Uneducated, But Ambitious and Self-Absorbed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Kim Jong-il: Spoiled and Ambitious Young Man Seeks Absolute Power 121 Chapter 9. Conclusion 126 Appendix: Additional Hypotheses Testing 130 Collapsing Constitutional and Nonconstitutional Power Successors: All Power Successors vs. Power Seizers . . . . . . . . . . . . . . . . . . . . . . . 130 Hypothesis: Power Successors vs. Democracies . . . . . . . . . . . . . . . . 133 Distinguishing Power Seizers and Comparing with Leaders Under Similar Constraints : Power Seizers (hereditary) vs. Similar Constraints; Power Seizers (nonhereditary) vs. Similar Constraints . . . . . . . . . . . . . 135 Distinguishing Power Seizers and Comparing with Democracies . . . . . . . 138 Appendix Table 14: List of Power Seizers 143 Appendix Figure 11: World Map by Authoritarian Leader 150 Appendix Table 15: List of Power Successors (Constitutional) 151 Appendix Figure 12: Weeks’ 2012 Militarism and Personalism Index 153 Appendix Figure 13: GWF 2014 Classification of the nine nonconstitutional Power Successor countries 154 Appendix Table 16, 17, 18, 19: List of Disputes Under Chiang Kai-shek, Chiang Ching-kuo, Kim Il-Sung, Kim Jong-il 155 References 163 4 List of Figures 1 Number of MIDs/Years in Office by Power Seizers and Constitu- tional Power Successors . . . . . . . . . . . . . . . . . . . . . . . . . 11 2 Operationalization of Autocrats . . . . . . . . . . . . . . . . . . . . 27 3 Distribution of XCONST Within Each Group: Figures indicate the mean value 1 standard deviation . . . . . . . . . . . . . . . . . . 38 4 Change of XCONST Over Time in the Power Successors (Constitu- tional) Compared to Democracies and Other Non-Democracies . . 39 5 Change of XCONST in the Nonconstitutional Dynasties Across the Hereditary Power Seizers and the Power Successors (Nonconstitu- tional) Over Time) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 6 Expected Percentage of Conflict Initiation: Iraq 1990 Scenario . . . 55 7 Expected Percentage of Conflict Initiation Compared to Democracy: Iraq 1990 Scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 8 Number of MIDs Divided by Years in Office by Leaders . . . . . . . 70 9 Expected Percentage of the Time That a Country Will Initiate Con- flict: Iraq 1990 Scenario (Using CLARIFY) . . . . . . . . . . . . . . 139 10 Expected Percentage of the Time That a Country Will Initiate Con- flict: Iraq 1990 Scenario . . . . . . . . . . . . . . . . . . . . . . . . . 141 11 World Map by Authoritarian Leader . . . . . . . . . . . . . . . . . 150 12 Weeks’ Militarism and Personalism Index in the nonconstitutional Power Successor countries. . . . . . . . . . . . . . . . . . . . . . . . 153 13 GWF Classification of the nine nonconstitutional Power Successor countries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 5 List of Tables 1 Descriptive Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2 Nonconstitutional Dynasties with the First Generation Power Seiz- ers and the Succeeding Power Successors . . . . . . . . . . . . . . . 15 3 MID Initiation by Power Seizers vs. Power Successors . . . . . . . . 53 4 Power Seizers and Power Successors vs. Democracies . . . . . . . . 57 5 MID Initiation by Power Seizers vs. Autocrats Under Similar Insti- tutional Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 6 MID Initiation by Power Successors vs. Autocrats under Similar Institutional Constraints . . . . . . . . . . . . . . . . . . . . . . . . 64 7 MID Initiation by Power Seizers vs. Power Successors . . . . . . . . 68 8 Breakdown of MIDs Based on Intensity in North Korea . . . . . . . 77 9 Breakdown of MIDs Based on Intensity in Taiwan . . . . . . . . . . 78 10 MID Initiation by the Power Seizers vs. Collapsed Power Successors 132 11 MID Initiation by the Collapsed Power Successors Compared to Democracies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 12 MID Initiation by Power Seizers (when distinguished between "hered- itary" and "nonhereditary") vs. Leaders Similar Institutional Con- straints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 13 Power Seizers vs. Power Successors vs. Democracies . . . . . . . . 142 14 List of Power Seizers . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 15 List of Power Successors (Constitutional . . . . . . . . . . . . . . . . 151 16 List of Disputes under Chiang Kai-Shek (1950-1974) . . . . . . . . . 155 17 Disputes Under Chiang Ching-kuo (1978-1987) . . . . . . . . . . . . 156 18 Disputes Under Kim Il-Sung (1946-1993) . . . . . . . . . . . . . . . 156 19 Disputes Under Kim Jong-il(1994-2010) . . . . . . . . . . . . . . . . 158 6 Chapter 1: Introduction Isa Ibn Al-Khalifah of Bahrain, Jigme Dorji Wangchuk of Bhutan, Hussein Ibn Talal El-Hashim of Jordan, and Jean Duvalier of Haiti – These are some names not com- monly associated with international conflict. Isa Ibn Al-Khalifah, the Emir of Bahrain from 1971 to 1998, for example, forged a pro-Western line in foreign policy hosting the United States Fifth Fleet in the Gulf, while maintaining an alliance with the region’s economic and military powerhouse, such as Saudi Arabia. When he died of a heart at- tack in 1998, then Secretary-General of UN, Kofi Annan, described him as “force for stability in troubled region”. 1 Jigme Dorji Wangchuk, the King of Bhutan from 1952 to 1971 strived to end the country’s decade-long isolation in the international community and joined the United Nations in 1971 while heralding economic and social reforms into the country’s quasi-feudal system. King Hussein of Jordan (1952-1999) left the legacy of securing a peace with Israel in 1994, a stunning achievement given the tumultuous politics of the Middle East then. At the Wye River Summit in 1998 initiated by then US President Bill Clinton, King Hussein declared that there had been “enough destruc- tion, enough death, enough waste” during the five decades of Arab-Israeli conflict, and that “we have no right to dictate through irresponsible action or narrow-mindedness the future of our children or their children’s children.” 2 Even Jean Duvalier, whose brutal crackdown on domestic opponents and notoriously lavish spending out of palace coffers 1 United Nations. 1999. Secretary General Expresses Sadness at Death of Bahrain’s Amir, Describes Him as ‘Force of Stability’ in Troubled Region. New York:UN Press Releases. Available at <http://www.un.org/press/en/1999/19990308.sgsm6918.html>. Accessed in February 2017. 2 Judith. February 8, 1999. Death of a King; Cautious King Took Risks In Straddling Two Worlds.New York Times. Available at <http://www.nytimes.com/learning/general/ onthisday/bday/1114.html> Accessed in February 2017. 7 in an impoverished nation earned him international contempt and derision had not been known to be particularly violent in his foreign policy pursuit. 3 It may then come as a puzzle to learn that these leaders enjoyed as firm a grip over their country’s domestic political life as Idi Amin of Uganda, Saddam Hussein of Iraq, and Kim Il-Sung of North Korea 4 whose names we often take as synonymous with international conflict and violent foreign policy (Weeks 2012). For example, Emir Isa Ibn Al-Khalifah dissolved the Parliament in 1975 suspending the nation’s Constitu- tion, and finally in 1992 replaced the National Assembly with a 30-member consultative council whose members would be appointed by the Emir. Jean Duvalier relied on a bru- tal militia known as the Tontons Macoutes (Creole for "bogeymen") that left thousands of his own people dead, disappeared or illegally detained. What explains the relative peace with the outside world by these personalist dictators? The existing IR literature with a focus on domestic institutional constraints on leaders suggests that a leader’s level of personalism, the degree to which a leader per- sonally controls the state apparatus without facing a strong domestic audience (Chehabi and Linz 1998; Geddes 2003; Weeks 2012), is an important indicator for the regime’s 3 According to the Militarized Interstate Disputes (MID) dataset (Palmer, d’Orazio, Kenwick, and Lane, 2015), there is a record of only one dispute initiation by Emir Isa Ibn Al-Khalifah (1971-1998); none by Kimg Jigme Dorji Wangchuk (1952-1971); none by Jean Duvalier (1971-1985); and eleven MIDs by King Hussein (1952-1998). 4 According to Polity IV Dataset (Marshall, Jagger, and Gurr, 2009), Bharain under the Emir’s rule records a XCONST score of 1, POLITY2 score of -10 except for the Emir’s final years (XCONST: 2, POLITY2: -9); Bhutan a XCONST score of 1 and POLITY2 score of -10 under King Jigme Dorji Wangchuck; Jordan a XCONST score ranging from 1 to 3 and a few occasional years with 5, and POLITY2 score ranging from -10 to -9 with -4 and -2 in King’s final years; Haiti under Jean Duvalier a constant 1 for XCONST, and a POLITY2 score ranging from -10 to -9. For more information on XCONST, see the “Operationalization” section. 8 conflict propensity with other countries. A personalist leader whose supremacy in do- mestic affairs is often enabled by forceful means such as torture and killing of his own citizens not only is free to set foreign policy goals at his whim, but face few domestic consequences for defeat or for starting foolish wars. As a result, personalist dictators are more likely to initiate military disputes than nonpersonalist leaders(Weeks 2012). Yet, are all personalist leaders equally belligerent internationally? As much as some personalist leaders, such as Kim Il-Sung, Saddam Hussein, and Idi Amin had indeed pursued violent foreign policies, others, such as Jean Duvalier, did not, and yet others, such as Emir Isa Ibn Al-Khalifah, had actively pursued peace. What makes some personalist leaders more prone to conflict than others? If not the level of personalism as suggested by the large literature, to what can we attribute this variation in the likelihood of conflict initiation amongst different personalist leaders? My dissertation shows that not the domestic institutional constraints on the leader, but a leader’s predisposition towards international violence better explains the initiation of international military conflicts by autocrats. I argue that not all personalist dictators who are insulated from strong domestic audience pursue reckless, violent for- eign policy path as the likes of Kim Il-Sung or Saddam Hussein have us believe. Instead, I offer an alternative measure to understand violent behaviors by autocrats: self-selection into power. More specifically, I argue that autocrats who rose through the ranks (i.e. se- lected themselves into power) typically by forceful means such as coup or assassination tend to be more belligerent than other dictators under similar constraints. On the con- trary, autocratic leaders whose absolute power over domestic affairs were granted on them by birth right (i.e. did not select themselves into power) are less likely to engage in militarized disputes with other countries compared to other dictators even under sim- ilar constraints. Jean Duvalier, for example, is famously known for being unwilling to take the helm of his country as a young heir, rather wanting either his elder sister or 9 his mother to take the crown. Moreover, this dissertation shows that these hereditary autocrats are not only less likely to initiate disputes compared to the personalist dicta- tors who rose via forceful means, but are also no more belligerent internationally even when compared to the democratic leaders. In sum, not all leaders exploit the concentra- tion of power accumulated on their hands to risk violent foreign policy behavior even when they can. The typical violent authoritarian leaders are those who come to power as a result of use of force, such as coups, assassinations, and revolts. However, those autocrats whose power acquisition came through hereditary ascription are significantly less conflict-prone than their first generation of autocratic ancestors even under similar institutional conditions. Table 1 and Figure 1 illustrate descriptive statistics of the number of MIDs by each leader type. Since the number of MIDs is divided by the length of office by the leader, the statistics provide a more accurate picture of conflict tendencies by different types of leaders independent of the effects of length in power. My theory predicts that, on average, the Power Seizers are involved in more conflicts than the Power Successors. Table 1 and Figure 1 confirm this prediction. Table 1: Descriptive Statistics Leader Type Mean SD Min Max N MID/Years (Power Seizer) 0.26 0.56 0 6.2 280 MID/Years (Power Successors, Constitutional) 0.13 0.2 0 1 58 MID/Years (Power Successors, Nonconstitutional) 0.15 0.2 0 0.65 10 To demonstrate my argument, I seek to bring hereditary regimes to light, an au- thoritarian regime type where the rulers retain personalist power throughout generational leadership transitions (see Figure 4 and 5), while the mode of succession (i.e. heredi- tary succession) ensures relatively peaceful transition unlike in other typical personalist 10 Figure 1: Number of MIDs/Years in Office by Power Seizers and Constitutional Power Successors 11 regimes where power acquisition usually occurs through forceful means. I show that the succeeding monarchs in hereditary regimes who are arguably as personalist (Wahman, Teorell and Hadenius 2013), yet, who had not sought their supreme domestic power by force, tend to be less conflict prone on the international stage compared to other auto- crats who used violence to obtain domestic power. In sum, a personalist leader, defined as a powerful ruler who dominates the government apparatus, and is not constrained by institutional alignments (Weeks 2012), tends to be less belligerent abroad if he came to power through relatively peaceful means. After a critical examination of the current literature on authoritarian regimes and international conflict in Chapter 2, Chapter 3 puts forth a theoretical framework that highlights two questions crucial to understanding conflict behavior by autocrats (see Figure 2). First, did the autocratic leader “select himself into power”? I call those who had managed to rise through the ranks to seize power through forceful means such as coup or revolts the Power Seizers, and those whose power acquisition is determined by birth right and hereditary ascription the Power Successors. While competition and contestation for power among potential heirs may always exist (Herb 1999), I assume that the level of self selection among candidates by birth right – the succeeding gener- ations – is marginal at best compared to that of leaders who chose to fight for power from humble background against political establishment. Previous research has abun- dantly shown that the tumultuous path towards becoming a personalist dictator requires a person who is ambitious and views force as effective (Colgan and Weeks 2015; Gurr 1988; Skocpol 1979). The Power Seizers who successfully managed to confront the challenges of attaining and maintaining power typically regard using military force as effective, appropriate, even necessary, resulting in recourse to violence even on the inter- national stage (Colgan 2010; Gurr 1988; Horowitz and Stam 2014; Colgan and Weeks 2015). In particular, research on revolutionary leaders concurs that the types of individ- 12 uals who succeed as revolutionaries are more risk tolerant and ambitious than typical leaders (Colgan 2012). These traits in turn make it more likely that the leader will instigate international conflicts (Colgan and Weeks 2015; Gurr 1988; Skocpol 1979). On the contrary, if we follow the same logic that the tumultuous path towards securing a personalist regime selects for a risk tolerant and ambitious leader, the Power Successors may lack those characteristics precisely because they didn’t fight for the po- sition of paramount power they find themselves in, but rather simply assumed their po- sition by birth right. The lack of grand ambition and risk tolerance of those succeeding leaders relative to the Power Seizers make it less likely that they would instigate inter- national conflicts at the same rate as the Power Seizers would. Moreover, the process of attaining power forgoing fierce and sometimes bloody competition may have limited the chances for the Power Successors to become accustomed to the belief that violence is an effective policy option. As a result, the Power Successors are less likely to rely on force to resolve international disputes than the Power Seizers not due to varying lev- els of institutional constraints, but as a function of predisposition towards international violence. The second question involves distinguishing the “nonconstitutional dynasties” such as North Korea and Syria (Table 2) from the “constitutional monarchies” such as Saudi Arabia and Kuwait among the Power Successors. Both the “constitutional” as well as the “nonconstitutional” Power Successors share the trait of assuming power by hereditary succession. However, the paths to power by those “nonconstitutional” Power Successors are believed to be more challenging whose succession is not a constitution- ally established practice in these countries. Unlike their counterparts in “constitutional” monarchies, heirs in those republic-style, ”nonconstitutional” dynasties may face more challenges and threats to their authority from potential political enemies, which in turn will make them more likely to resort to violence in their consolidation of power. Kim 13 Jung-un, for example, is believed to have personally ordered the assassination of his half brother, Kim Jong-nam, who had publicly criticized the third generation dynastic succession in North Korea. 5 Thus, while both “constitutional” as well as “nonconsti- tutional” Power Successors are expected to be less likely to instigate international con- flicts compared to the Power Seizers, the extent to which the rate of dispute initiation is depressed may differ between these two groups of successors. I expect the “nonconsti- tutional” Power Successors to demonstrate more international violence. Chapter 4 explores these distinctions of authoritarian leaders through opera- tionalization of Power Seizers and Power Successors based on the criteria of how they came to power using a newly compiled dataset that relies on some of the most widely used datasets by the scholars in the field. 6 5 Fifield, Anna. For Kim Jong Nam, a sad ending to a lonely life. The Wash- ington Post. February 24, 2017. Available at <https://www.washingtonpost.com/ world/asia_pacific/the-sad-life-and-even-sadder-end-of-kim-jong-nam/2017/02/24/ 7fb591fc-f941-11e6-aa1e5f735ee31334_story.html?utm_term=.2c8cb16674e3>. Accessed in February 2017. 6 That includes Colgan’s Revolutionary Leader Dataset (Colgan and Weeks 2015), Polity IV Dataset (Marshall, Jagger and Gurr 2009) and Brownlee’s data (Brownlee 2007). 14 Table 2: Nonconstitutional Dynasties with the First Generation Power Seizers and the Succeeding Power Successors Azerbaijan Dominican Republic Haiti Nicaragua North Korea Singapore Syria Taiwan Togo First Heydar Rafael Francois Anastasio Somoza Kim Lee Hafez Chiang Gnassingbe Generation Aliyev Trujillo Duvalier Garcia Il-sung Kuan Yew al-Assad Kai-shek Eyadema Power Seizers (r.1993-2003) (r.1930-61) (r.1957-71) (r.1936-56) (r.1948-94) (r.1965-2004) (r.1971-2000) (r.1949-75) (r.1967-2005) Nonconstitutional Ilham Rafael Jean-Claude Luis Somoza Kim Lee Bashar Chiang Ganssingbe Power Successors Aliyev “Ramfis” Duvalier Debayle Jong-il Hsein Loong al-Assad Ching-kuo Faure (r.2003-) Trujillo Martinez (r.1971-86) (r.1956-67) (r.1994-2012) (r.2004-) (r.2000-) (r.1975-88) (r.2005-) (r.1961) Anastasio Somoza Kim Debayle Jung-un (r.1967-1979) (2012-present) 15 Chapter 5 carries out extensive statistical analyses using the dataset described in Chapter 4. The results provide strong support for my argument: The Power Succes- sors are less likely to initiate international conflicts than the Power Seizers under similar institutional constraints. Moreover, the “constitutional” successors, in particular, tend to be even more peaceful than democratic leaders when it comes to international conflict initiation. The subsequent two chapters turn to two case studies: North Korea as a de- viant case, and Taiwan as a typical case. By closely examining those two cases, this dissertation hopes to offer a fruitful integration of quantitative techniques laid out in Chapter 5 and qualitative techniques, a line of inquiry suggested by a number of recent studies (Bennett and George 2005; Brady, Collier and Seawright 2004; Gerring 2001, 2007; Goertz 2006; King, Keohane and Verba 1994; Ragin 2000; Levy 2008). The final chapter of this dissertation summarizes implications for the study of conflict initiation by authoritarian regimes and attempts to offer some valuable insights to foreign policy analysts facing autocrats abroad. This dissertation combines insights from the study of authoritarianism in com- parative politics with those in IR regarding conflict initiation by autocrats. In particular, I seek to bring monarchies to light, an authoritarian regime type that has long been underexplored by scholars forgoing the same level of investigation into the regime’s do- mestic institutional constraints other authoritarian regime types are typically subject to. For example, monarchies were added only recently to Geddes’ updated authoritarian regime typology dataset (2014) 7 after its initial omission in her earlier data (1999a), and have not been part of research question regarding conflict initiation by autocrats by some 7 See Appendix Figure 13 16 of the most prominent scholars in the field. 8 .The obscure classification of monarchies perhaps reflects the notion throughout the history of political thought that monarchs are somehow distinct from tyrants (Cheibub, Gandhi and Vreeland 2010), which is ev- ident in most of the recent authoritarian regime typology datasets where monarchies are marked as a separate group, yet are still included in the sample of dictatorships. 9 Although monarchies may indeed have qualities distinct from other autocratic regime types such as strong reliance on family(Herb 1999), they are considerably more similar to each other in that the rulers are unconstrained by their domestic audience as well as institutions, which is the main functional aspect I focus on here to classify as a person- alist regime. I later show that the XCONST variable from Polity IV Dataset measures the "extent of institutionalized constraints on the decision-making powers of chief ex- ecutives"(Marshall 1999, p.24) in hereditary regimes as similar, if not less, as in other authoritarian regimes (Figure 3, 4, 5) confirming that monarchies are as personalist as other dictatorships. This dissertation attempts to differ from the existing research in that it includes hereditary regimes in the analysis of conflict initiation by authoritarian regimes, instead of treating monarchies as a distinct group operating under some myste- rious essence of kingship. I follow Wahman and his colleagues’ insight that "monarchies are by definition the most personalist regime type possible" and "they share many char- acteristics with the personalist regimes” (Wahman, Teorell and Hadenius 2013, p.25). I intend to hold up monarchies against the same criteria as other authoritarian regimes to conduct a comprehensive analysis on the relationship between domestic constraints and 8 See for example, Weeks (2012); Peceny, Beer and Sanchez-Terry (2002); Peceny and Beer (2003); Peceny and Butler (2004) 9 See for example, Geddes, Wright and Frantz (2014); Cheibub, Gandhi and Vreeland (2010); Wahman, Teorell and Hadenius (2013) 17 conflict initiation in authoritarian regimes. Furthermore, this dissertation highlights the nontraditional dynasties whose rate of expansion averages to one every three years since the installation of North Korea’s Kim Jong-Il in 1994 (Brownlee 2007), 10 and shed light on how their conflict propensity compares to that of traditional constitutional monarchs as well as other types of autocracies. The inclusion of these non-constitutional dynasties in “republics” serves as a reminder to the study of authoritarianism that extension of absolute power through dynastic succession in fact occurs routinely outside the Arab world, 11 and that heredi- tary succession should not be portrayed as some cultural fluke unique to the Middle East (Lewis 2000). The findings have many implications for the study of conflict initiation by authoritarian regimes and offer valuable insights to foreign policy analysts facing autocrats abroad. 10 For example, hereditary succession in Syria in 2000; Azerbaijan in 2003; Singapore in 2004; Togo in 2005; and again in North Korea in 2012. Note that Mubarak in Egypt, Gadhafi in Libya, and Saleh in Yemen were also on their paths to groom their sons to rule until the uprising during the Arab Spring stopped them. 11 On the surface, the notion that monarchies are a product of the Arab world seems to carry weight considering that only a handful of absolutist kingships remain outside the Arab world, such as those in Swaziland and Brunei after the last king of Bhutan gave up his supremacy in favor of parliamentary democracy in 2008. 18 Chapter 2: Literature Review: Domestic Institutional Con- straints and International Conflict Initiation by Authori- tarian Leaders I combine insights from two lines of large literature in IR and comparative politics to better understand international conflict initiation by autocrats. In particular, my argument builds on the growing literature on conflict behavior by autocrats in IR as well as the mounting interest in monarchies put center stage in comparative politics especially after the Arab Spring in 2011. Despite the vast amount of scholarly attention to authoritarian regimes and their conflict propensity, there is scant research explor- ing the relationship between monarchies and their foreign policy paths. IR scholarship typically excludes monarchies from a representative sample of autocrats, whereas the interest of comparative studies on monarchs has fallen short of the regimes’ conflict be- havior. Moreover, the excessive focus on domestic institutions in the study of autocratic regimes in IR fails to account for individual variation across leaderships under similar constraints. In some cases where leaders’ perceptions and predispositions have been addressed, the effects of institutions and leaders are often conflated making it hard to assess them individually. First, existing findings on autocratic conflict behavior in IR explicitly omit monarchies in the sample of dictatorships (Weeks 2012; Peceny, Beer and Sanchez- Terry 2002; Peceny and Beer 2003; Peceny and Butler 2004), and in cases they do incor- porate monarchies, the inclusion lacks specific theoretical underpinning, and typically goes without notice (Lai and Slater 2006). As a result, we know very little about conflict behavior by monarchs and how it compares to other types of authoritarian regimes and democracies. The same problem persists in comparative politics where, for too long, monarchs had been treated as a cultural curiosity of Arab politics and an autocratic in- 19 stitution that would not long exist. Huntington(1968) famously coined the term “King’s Dilemma” in which he alleged the end of monarchs through the contradiction between the monarchy as a quintessential traditional institution and the challenges of modern state-building that requires kings to share authority with crucial new groups such as the middle class. 12 However, contrary to then widely held notion that monarchies are passé (Anderson 1991), monarchs demonstrated a remarkable resiliency and durability in the wake of Arab Spring in which all eight 13 monarchies in the Middle East survived the mass demonstrations and uprisings other republics 14 in the region gave in. The sustain- ability of the Middle Eastern monarchs has sparked a renewed interest with scholars exploring different causes for the monarchical stability since then. Some of the central factors include the political culture that provides legitimacy to the region’s monarchs (Menaldo 2012; Schlumberger 2010); military support of foreign patrons such as the United States, or historically, United Kingdom, France and the Soviet Union (Brand 2013; Yom 2013; Yom and Gause III 2012) ; continuous and high rent influx via oil and natural gas export (Gause 1994; Yom and Gause III 2012; Luciani 1990; Bellin 2012); intra-family ruling institution (Herb 1999) ; and the presence of coercive appa- ratus (Bellin 2012, 2004; Todd Spinks, Sahliyeh and Calfano 2008). Although recent scholarly inquiry into monarchs has started to open up room to explore the internal logic and operation within monarchical regimes, the dependent variable in question has been 12 For more on the “End of Monarchies,” see Burling (1974); Herz (1952); Bienen and van de Walle (1991); Anderson (1991); de Mesquita, Smith, Siverson, Morrow (2003). 13 There was little to no protest in Qatar, United Arab Emirates, Saudi Arabia, and Oman; a significant level of protests but moderate demands in Morocco, Jordan, and Kuwait. Bahrain was the only monarchy that saw large protests, and even there the protesters at first did not seek regime change as they had in the republics. 14 Tunisia, Libya, Egypt, and Yemen saw regime change. 20 almost entirely fixated on the regimes’ sustainability 15 evading a multitude of other po- tential questions such as their foreign policy choices. As the field stands today, we are starting to know about how and why monarchies persist in the Middle East, but we know little about their conflict behavior in comparison to other authoritarian regimes. The second problem lies in that the existing literature on monarchy mainly ex- plores traditional, constitutional monarchs in the Middle East only, forgoing a number of nontraditional hereditary regimes where power succession from father to son is not a constitutionally accepted practice, yet nevertheless is a defining element of the regime’s leadership configuration. The vast majority of research focuses on the eight constitu- tional monarchies in the Middle East. 16 As a result, dynasticism in republics, such as Syria and North Korea are largely ignored in the study of monarchies even though their regulation of power succession bears most semblance to constitutional monarchs than any other types of regimes. Similarly, the exclusive focus on the Middle East has re- sulted in neglecting hereditary monarchies outside the Arab world, such as Swaziland and Brunei, and historically countries like Bhutan and Cambodia. Brownlee’s (2007) in- sight into dynasticism in modern autocracies is a unique contribution that broadens the scope of hereditary succession across unconventional monarchies. His sample of hered- itary dynasties incorporates geographical expansion beyond the Middle East as well as across time periods to include countries such as Singapore and Taiwan, not commonly associated with monarchical regimes. 15 As an exception, Menaldo (2012), for example, employs Banks’ Conflict Index as the dependent variable, which is a weighted average of Assassinations, Strikes, Guer- rilla Warfare, Government Crises, Purges, Riots, Revolutions, and Anti-Government Demonstrations. The purpose here, too, however, was to explore regime stability of the Middle Eastern monarchs. 16 Most literature exploring the monarch sustainability belongs here. 21 Third, previous research reveals little about the extent and contents about insti- tutional constraints on the monarchs in terms that are comparable to other authoritarian regimes. In a broad sense, two conflicting accounts dominate the discourse revolving around the extent of king’s authority in monarchies. On the one hand, the conventional view that kings’ power is absolute, and that kings organizationally stand above everyday politics is used to explain how monarchs persist. Kings are able to initiate economic and political reforms in response to popular pressure with fewer constraints than their coun- terparts in republics who are encumbered by the interests of other domestic factions, such as military, the ruling party and bureaucrats (Lucas 2004). An alternative perspec- tive suggests that the practice of putting kin in key state offices helps enhance regime unity by preventing elite defections (Herb 1999), but it also means considerable con- straints on the king in embarking on new political initiatives. In particular, Kings in the five dynastic monarchies of the Persian Gulf 17 where the royal family forms a ruling in- stitution cannot dismiss his relatives from their posts, and the family even retains the au- thority to remove the monarch to replace him with another member of the family (Herb 1999, p.238). 18 Either account, however, eludes measures that are operationally com- parable to other political regimes in a way that allows researchers to meaningfully jux- tapose the effects of institutional constraints on the executives across different regimes. This same problem endures in almost all currently available regime typology datasets where monarchies are quickly set aside as a distinct category based on the single criteria that power succession is hereditary. This practice in effect forecloses the opportunity 17 Saudi Arabia, Kuwait, Bahrain, Qatar, United Arab Emirates. 18 Herb distinguishes between dynastic monarchies where the royal family forms a ruling institution and non-dynastic monarchies in which the king rules alone. Kings of non-dynastic regimes such as Shah of Iran exert absolute control over state apparatus. 22 of investigating monarchies against the same institutional measurement as other regime types. For example, for Cheibub, Gandhi, and Vreeland (2010), monarchs are classified as such if the effective head “bears the title of “king” and has a hereditary successor and/or predecessor.” In other words, our conceptualization of monarchies has largely been definitional (i.e. hereditary succession), rather than instrumental (i.e. what kind of domestic institutional constraints, if any, are kings subject to?), which lends little use in comparing monarchies with other autocratic regimes. Lastly, the vast majority of attention in IR scholarship on conflict initiation has been directed to the domestic institutional constraints on the executive decisions evading questions on individual variation under similar constraints. A series of earlier studies are centered around the different conflict propensity between democratic and authoritarian political systems (Fearon 1994; Reiter and Stam 2002; Bueno de Mesquita et al. 2003). More recently, scholars have started to explore the different strains within the authori- tarian regimes and emphasized the variation across the different types inside the broad category of autocracies (Geddes 2003; Weeks 2008). This line of research mostly draws on two central dimensions to understand the internal logic of authoritarian regimes: first, whether the leadership is civilian or military; and second, the extent of personal power, or personalism, of the leader (Geddes 2003; Weeks 2012). 19 This two-dimension typol- ogy of authoritarian regimes conceptually relies on the ex post constraint on the leader - That is, the ex post accountability leaders have to deal with when they deliver outcomes that are deemed unacceptable, such as losing a war. On the one hand, the level of per- sonalism of a regime speaks to both the degree of which a dictator can freely choose a policy goal and pursue it (ex ante constraint), and also to how vulnerable he or she 19 For alternative ways of differentiating among authoritarian regimes, see, for ex- ample, Danilovic and Clare (2004); Cheibub, Gandhi and Vreeland (2010); Ezrow and Frantz (2011); Gandhi and Przeworski (2006); Teorell and Hadenius (2007). 23 might be to the cost of defeat after the fact (ex post constraint). 20 Unconstrained person- alist dictators are less likely to be severely punished even in the aftermath of foreign policy disaster, and thus disincentivized to pay greater caution in fighting wars. On the other hand, the military-civilian dimension reveals the "preferences" and perceptions of the leader and the domestic audience 21 towards using coercive force abroad (Weeks 2012). The logic is that the leader and the domestic audience with military background typically have a greater desire and tolerance for the initiation of military conflict(Sechser 2004; Horowitz and Stam 2014). On the contrary, civilian leaders and domestic audi- ence typically dread the prospect of military disaster abroad and thus are less likely to initiate militarized dispute. While the literature based on two-dimension typology partly does address in- dividual variation towards using military force, it is not clear to what extent we can attribute those effects either to the individual or to the institution. For example, if a personalist leader is more conflict-prone, is it because of his personal belligerency that thrives in such an unconstrained regime type (Colgan and Weeks 2015; Weeks 2012), or is it because of the institutional failure to hold the leader accountable (Geddes 2003; Weeks 2012)? Similarly, if a military leader is more likely to initiate military disputes abroad, is it because of his predisposed view that using military force is effective and necessary (Colgan 2010; Gurr 1988; Horowitz and Stam 2014; Weeks 2012), or is it 20 Some research argues, however, that although personalist dictators are less likely to be punished than nonpersonalist dictators, once they are, they are more likely to face severe punishment such as exile or death than their nonpersonalist counterparts who more face a quiet retirement (Debs and Goemans 2010). 21 For different interpretation of military-civilian dimension, see, for example, Lai and Slater (2006), who focus on the source of infrastructural power in the military vs. civilian regimes. 24 because the domestic institution dominated by the military tends to be more forgiving of an unsuccessful military campaign (Posen 1984; Snyder 1984; Weeks 2012)? The scholarship conflates the effects of institutions and leaders 22 making it hard to assess individual effect independent from the other. In sum, my argument draws on the personalism dimension of the existing typology, but focuses on the significance of the individual predisposition and perception rather than institutions. With the growing recognition that individual leaders matter for a country’s foreign policy path (Byman and Pollack 2001; Saunders 2011), scholars have started to explore variation in the conflict behavior of autocrats (Colgan and Weeks 2015). 23 My theory builds on this growing literature. Furthermore, I attempt to bridge the gap between conflict studies in IR and research into monarchies in comparative politics and form a comprehensive analysis on conflict behaviors by autocrats. 22 The term “personalism” I used in the introduction of this dissertation rather refers to the institutional aspect of authoritarian regimes. 23 More recent work has started to actively integrate research from social psychol- ogy in interdisciplinary manner and explores the relationship between foreign policy behavior and dispositional traits such as time preferences (Kreps 2008; Frederick 2006; Kertzer 2013) ; risk preferences (McDermott 2001; Kertzer 2013) ; culture of honor (Tetlock 2003; Ginges and Atran 2011; Kertzer 2013) , trait self-control (de Ridder et al. 2012; Kertzer 2013); prosocial vs. proself (Rathbun, Kertzer and Paradis 2014) ; and moral values (Kertzer et al. 2014). 25 Chapter 3: A Theoretical Framework on Leadership Type and Conflict Initiation: A Succession-Based Explanation Three Main Hypotheses: Power Seizers vs. Power Successors vs. Democratic Leaders In my theoretical framework, self-selection into power is the key distinguish- ing feature between the Power Seizers and the Power Successors. I argue that the means through which personalist leaders came to power explains the level of conflict-proneness during their reign. I define personalist leader as a powerful ruler who dominates the gov- ernment apparatus, and is not constrained by institutional alignments, such as effective politburo, party hierarchy, merit-based promotions, and rules ensuring the turnover of leaders (Weeks 2008). Existing IR literature has us predict that those unconstrained per- sonalist leaders tend to be invariably conflict-prone compared to nonpersonalist leaders. This dissertation shows that not all unconstrained persoanlist leaders pursue belligerent foreign policy if we expand our datasaet to include hereditary successors (who are per- sonalist leaders) and distinguish them based on how they came to power (See Figure 2). Focusing on the selection process of personalist leaders into power reveals two mechanisms that explain why the Power Seizers are more likely to initiate conflicts than the Power Successors. The first mechanism involves disposition and traits inherent in the leaders themselves. Previous research has shown that personalist leaders who came to power through forceful means tend to have greater international ambitions and partic- ularly revisionist views (Weeks 2012; Colgan and Weeks 2015). A number of large-N studies support that forceful leaders, once in power, are more likely to reflect their grand ambitions in the pursuit of foreign policy paths and will be less reluctant in resorting to 26 Figure 2: Operationalization of Autocrats military force in the international arena if deemed necessary (Colgan and Weeks 2015; Colgan 2013a,b). Describing individuals who succeed as revolutionary leaders as risk tolerant and ambitious than typical leaders, Colgan argues that those characteristics help leaders to alter the status quo not only domestically but also in the international realm (Colgan 2012). Following the same logic, I argue that the same personality traits, such as risk tolerance and grand ambition, which helped leaders to emerge and pull off in their domestic struggle also make them more likely to initiate international conflicts once they have obtained office. Gurr concurs that leaders who secured power through forceful means are inclined to respond with violence to future challenges even if those conflicts arise in the international arena (Gurr 1988). Moreover, the Power Seizers can be characterized as narcissist personalities who desire supremacy over others (Rosen 2009). Psychological studies on personalist dictators shed light on the insatiable need 27 for attention and dominance over others which shapes their grand ambition for inter- national conquest and leads the individuals to embark on perilous attempts to expand territory (Glad 2002). Saddam Hussein, for example, famously aspired to establish him- self as caliph with supreme power ruling over a pan-Arab caliphate created with his own hands. In sum, the first generation leaders tend to be more conflict-prone than typical leaders once in power because of the traits that also allowed them to succeed in their struggle to obtain domestic power. The Power Successors, on the other hand, are those who inherited upon their births the absolute power their first generation ancestors had built, and the domestic po- litical conditions that feed into it. They typically lack the grand international ambitions and revisionist world view that were the driving force for domestic dominance by their first generation leaders, precisely because they did not need them to obtain the supreme power in their hands. Since the authority was bestowed upon them by birth right, they did not have to fight for it. Jean Duvalier, for example, is famously known for being unwilling to take the helm of his country as a young heir, rather wanting either his el- der sister or his mother to take the crown. Absent these characteristics that make it more likely for a leader to seek international violence, the Power Successors tend to be less conflict-prone than the Power Seizers even when they are equally unconstrained by government instruments. The second mechanism involves the divergent experiences and perceptions about use of force leaders acquire in the process of obtaining power. Being exposed to violent, bloody struggle over the course of their success in claiming domestic power, the Power Seizers are likely to have developed practical skills necessary to command military forces under his rule. Skocpol, for example, argues that leaders who are suc- cessful in competition over domestic power become particularly skillful at organizing and mobilizing their people for mass campaign involving violence. Consequently, those 28 leaders tend have greater capacity for instigating conflict(Skocpol 1979). Furthermore, the Power Seizers who managed to survive the confrontations and challenges on their path towards power tend to be socialized into the belief that military force is useful and effective, and even regard force as a routine and legitimate instrument of politics (Posen 1984; Snyder 1984). Over time, this perception of military force as an appropriate tool can harden into offensive doctrine when dealing with dispute settlement internationally. Studies show that if violence was the primary means through which power was achieved and sustained, then violence tends to be applied “in all situations of stress, internal or external” (Brecher 1996, p.220). Thus, when confronted with an international chal- lenge, the Power Seizers are more willing to resort to military force than typical leaders because of their greater capacity to command force and their belief about its usefulness. On the contrary, the Power Successors are largely limited in their experiences with violence towards their path to power compared to the Power Seizers, and therefore, lack the skills to be put into practice when commanding military forces. Moreover, the relative lack of firsthand experiences with force means that they are less likely to have developed a belief about the usefulness and legitimacy about military force as a political tool. While competition and contestation for power among potential heirs may always exist (Herb 1999), I assume that the level of self selection among candidates by birth right is marginal at best compared to that of leaders who chose to fight for power from humble background against political establishment. Hypothesis: Power Seizers vs. Power Successors In sum, because of the divergent paths to power, the Power Seizers and the Power Suc- cessors respond differently to challenges that arise internationally. H1: The Power Seizers are more likely to initiate international conflicts than 29 the Power Successors under similar institutional constraints on the executives. How, then, do the Power Seizers and the Power Successors compare to demo- cratic leaders? The Power Seizers represent the typical personalist leaders who are reck- less and belligerent in their foreign policy pursuits (Weeks 2012). Democratic leaders, on the other hand, are elected officials chosen by the public through free and fair elec- tions on the premise that they would serve the best interests of the citizens during their term as defined by the Constitution. Although, arguably, democratic leaders also “se- lect themselves into power” as career politicians, the peaceful and lawful path towards power is not even remotely comparable to that typically experienced by the Power Seiz- ers. Thus, the Power Seizers have greater likelihood to initiate international conflicts than democratic leaders, a finding consistently supported by the large IR literature (Re- iter and Stam 2002; Fearon 1994; Bueno de Mesquita et al. 2003). The existing literature, however, is largely silent on the Power Successors in authoritarian regimes and how they would fare compared to democratic leaders in terms of their conflict propensity. In general, both the democratic leaders and the Power Suc- cessors experience relatively peaceful transitions of power compared to the Power Seiz- ers. Therefore, I first hypothesize that unlike the Power Seizers, the Power Successors are no more belligerent internationally than the democratic leaders. I further predict that the “nonconstitutional” Power Successors are more con- flict prone than the “constitutional” Power Successors. Although both the “nonconstitu- tional” and the “constitutional” Power Successors share the trait of power acquisition by birth right, there are reasons to believe that the path to power by the nonconstitutional Power Successors can be far more challenging than that by the constitutional Power Successors. The right to executive power of the succeeding heirs in countries like Saudi 30 Arabia and Kuwait is stipulated and protected by the Constitution. Hereditary succes- sion has long been an accepted tradition throughout many centuries in those monarchies that goes unchallenged by the public and, if any, by political oppositions. On the con- trary, the nonconstitutional Power Successors inherit republic-style autocracies such as North Korea and Syria where hereditary succession is neither constitutionally defined nor an accepted practice for a long period of time. Questions concerning legitimacy of those rulers from both the public and potential political enemies can challenge their path to power and the stability of leadership. It is more likely that the ”nonconstitutional” Power Successors will develop a sense of threat to his rule and take more precautionary measures in the process of consolidating their domestic power, which, in turn, is more likely to select for a belligerent leader. In sum, this discussion leads to the following hypotheses. Hypothesis: Power Seizers vs. Democratic Leaders H2: The Power Seizers are more likely to initiate international conflicts than democratic leaders. Hypothesis: Power Successors vs. Democratic Leaders H3: The Power Successors are no more likely to initiate international conflicts than democratic leaders. H3a: The ”nonconstitutional” Power Successors are more likely to initiate international conflicts than the ”constitutional” Power Successors. 31 Additional Hypotheses Testing: Power Seizers vs. Leaders under Similar Constraints, Power Successors vs. Leaders under Similar Constraints My main hypothesis is that the Power Seizers are more likely to initiate international conflicts than the Power Successors under similar institutional constraints due to their divergent personal characteristics and beliefs about military force. By extension, then, it may be conceivable that the Power Seizers are more belligerent not only compared to the Power Successors who may be particularly peaceful, but also compared to all other autocrats under similar institutional constraints. Similarly, the Power Successors should be less likely to initiate conflicts not only compared to the Power Seizers who are supposedly especially belligerent, but more peaceful compared to all the other autocrats facing similar institutional constraints. These two additional hypotheses pose a more rigorous test for my argument for individual predisposition and belief about force as opposed to institutional constraints. In order to isolate the effects caused by autocratic leaders’ pre-existing beliefs and personality traits, I further compare the conflict propensity of Power Seizers with that of all other autocratic leaders under similar constraints as the Power Seizers. If other autocratic leaders who similarly exert political dominance over state apparatus show significantly more reservation about using force, to the extent the Power Seizers are willing to use force, we can attribute the difference to the personal peculiarities of the first generations rather than to the institution. Hypothesis: Power Seizers vs. Autocrats Facing Similar Constraints H4: The Power Seizers are more likely to initiate international conflicts than all other autocrats facing similar institutional constraints. 32 To alienate the effect of personal predisposition of the Power Successors, I compare the rate of use of military force by Power Successors with that by all the other autocrats facing similar constraints. Again, to the extent that the rates differ, we can attribute the variation to the personal traits of the Power Successors rather than to the institutions. Hypothesis: Power Successors vs. Autocrats Facing Similar Constraints H5: The Power Successors are less likely to initiate international conflicts than all other autocrats facing similar institutional constraints. Lastly, although not specified in my main theoretical framework, one could further distinguish among the Power Seizers based on whether their descendants are the nonconstitutional Power Successors (Table 2). I call those Power Seizers, such as Kim Il-Sung, Hafez al-Assad, and Francois Dvualier, whose descendants are defined as nonconstitutional successors in my theoretical framework (see Figure 2), the “hered- itary” Power Seizers. The majority of Power Seizers who have failed to pass down power to their descendants are called the “nonhereditary” Power Seizers. 24 Theoreti- 24 The “nonhereditary” first generation Power Seizers include the typical personalist dictators such as Saddam Hussein and Idi Amin. Most first generation Power Seizers, however ambitious they may be, fail to pass down power to their descendants for reasons such as constraints by the party system, premature removal from power by opponents, or simply because they had no biological sons (Brownlee 2007). Another interesting point of debate is whether all dictators in fact wish to appoint their descendant as succes- sor. While the literature on dictatorship does not offer a systematic explanation for this question, it is largely noted that “president-for-life” is an aspiration for many despots (Geddes 1999b; Bueno de Mesquita et al. 2003) , and hereditary succession is a form of extension of corrupt and personalist autocracy (Linz and Stepan 1996). Tullock 1987 further argues that hereditary succession provides a method for regime stability during 33 cally, there should be no varying levels of conflict initiations between the “hereditary” and the “nonhereditary” first generations Power Seizers themselves. The main purpose of this distinction is to explicitly collect Nonconstitutional Dynasties which are made up of the “hereditary” Power Seizers and the “nonconstitutional Power Successors (Ta- ble 2). This distinction offers the merit of exploring any differences in the conflict propensity by the first generation, hereditary Power Seizers and the succeeding noncon- stitutional Power Successors in those Nonconstitutional Dynasties which happen to be the most belligerent regimes, such as North Korea. Within these nonconstitutional dy- nasties, I expect that the succeeding Power Successors are less conflict prone compared to the Power Seizers, their immediate predecessor. 25 Hypothesis: Hereditary Power Seizers vs. Nonconstitutional Power Successors H6: The Power Successors in nonconstitutional dynasties are less likely to initiate in- ternational conflicts than their immediate preceding Power Seizers. and beyond the ruler’s lifetime. If the ruler anoints a successor, this creates strong mo- tives for both assassinating him as well as reasonable security that he will get away with it. But failure to choose a clear successor is bound to increase the likelihood of an unregulated power struggle in the wake of the incumbent’s death. 25 For further additional hypothesis testing based on the hereditary and nonhereditary Power Seizers, see Appendix 10.3 and 10.4. 34 Chapter 4: Measurement and Operationalization of Main IVs and DV Measuring Domestic Institutional Constraints: Are the Power Seiz- ers and the Power Successors Equally Personalist Leaders? This dissertation challenges the existing rival hypothesis that the initiation of interna- tional conflict can be explained by the level of institutional constraints on the executive. Thus, before I compare the rates of conflict initiation between the Power Seizers and the Power Successors, I must first show that the level of constraints on these two groups of leaders is similar. Assessing these hypotheses requires data that measure the level and type of domestic constraints Power Seizers and Power Successors face. Previous attempts to capture this concept include Geddes typology (1999a; 2003) that distinguish between military, single-party, and personalist regimes 26 as well as Weeks (2012) authoritarian regime typology on machine, junta, boss, and strongman 27 . Weeks’ typology builds on Geddes raw regime data, and departs from earlier research in differentiating between military strongmen and civilian personalist dictators. She finds significant differences in conflict propensity between regimes led by the military strongman and civilian boss. Geddes, Wright, and Frantz’s (2014, GWF hereafter) autocratic regime data is an exten- sion and update of the earlier dataset by Geddes (1999a) with the additional coding of monarchy, which was absent in their earlier dataset. All these datasets, however, fail to provide information on the contents and level of domestic constraints faced by monar- chs, which is an essential component to test my hypotheses. In all cases, monarchies are coded as a distinct category alongside other authoritarian regime classification based 26 See Appendix Figure 13 27 See Appendix Figure 12 35 on familial succession, evading the same type of investigation other types of authoritar- ian regimes are subject to in terms of their domestic constraints on the executive. For example, monarchies in Weeks’ data (2012) are not coded in terms of her two main di- mensions – domestic constraints and preference of audience and the leader – and thus are not identified as either machine, junta, boss, or strongman. 28 Because monarchies are not held up against the same classification criteria as all the other authoritarian regimes are, we know little about the content and the degree to which domestic institution ex- erted constraints on monarchs in a meaningful way comparable to other authoritarian regimes. 29 Moreover, in some datasets, monarchies are differentiated from other au- thoritarian regimes in a way that conflicts with the core structure of monarchies. For example, for GWF, monarchies are those regimes where control over policy, leadership selection and the security apparatus is in the hands of a royal family (GWF, 2014, 8). If this is the case, monarchies can never be classified as personalist in the GWF data, although monarchies are by definition perhaps the most personalist regime type possible (HTW, 2013, 25 ). These coding practices perhaps reflect the notion held throughout the history of political thought that monarchs are somehow distinct from other tyrants. Yet, in institutional terms, monarchs belong to the wider sample of dictatorship because they 28 Weeks (2012) built independent measures of the two dimensions (personalism and military leadership) and created for every country both the indices representing the “de- gree” of the two dimensions as well as the category variables indicating either one of the four regime types: machine, boss, junta, strongman. Traditional monarchies as a whole elude this kind of investigation in her dataset, although some untraditional monarchs – most of which comprise the “nonconstitutional” Power Successors here– are subject to this classification. 29 It may be sensible to believe that monarchies are most similar to “boss” according to Weeks’s (2012) authoritarian typology in that the leadership in both groups is civilian (as opposed to military) and unconstrained (as opposed to constrained). 36 forgo competitive election (CGV 2010). To address this issue, I use the XCONST variable from Polity IV Dataset (Mar- shall, Jagger and Gurr 2009) which is ideal for assessing my hypotheses since all regime types, including the monarchies, are held up against the same measurement of executive constraints. XCONST is operationalized as the “extent of institutionalized constraints on the decision-making powers of chief executives” (Marshall, Jagger and Gurr 2009, 24). This variable concerns the checks and balances between the various parts of the decision-making process imposed by “accountability groups” such as the legislatures in Western democracies, the military in coup-prone polities, and the councils of nobles or powerful advisors in monarchies. Earlier research that used the XCONST variable to assess executive constraints includes Lai and Slater’s 2006 “despotic institution” that is coded as personalist if XCONST is either 1 or 2. A seven-category scale is used to capture the extent of executive constraints ranging from unlimited authority (coded as 1) to executive parity or subordination (coded as 7). 30 Based on the XCONST variable, I construct the extent of executive constraints for the Power Seizers and Power Successors. Figure 3 illustrates the distribution of XCONST in each category with the mean value and one standard deviation away from the mean. Not surprisingly, executive authority is highly unlimited across all the Power Seizers and the Power Successors (the mean value ranging from 1.77 to 2.08). Note that, if anything, the Power Successors have more free rein than their Power Seizers counterparts (i.e. the mean value for XCONST is lower for the Power Successors). This 30 1 (unlimited authority); 2 (intermediate category); 3 (slight to moderate limitation on executive authority); 4 (intermediate category); 5 (substantial limitations on execu- tive authority); 6 (intermediate category); 7 (executive parity or subordination). 37 Figure 3: Distribution of XCONST Within Each Group: Figures indicate the mean value 1 standard deviation confirms the notion by Wahman and his colleagues that monarchies are possibly the most personalist regime type (Wahman, Teorell and Hadenius 2013, 25). Based on the distribution of XCONST, I create a dummy variable called xconst- sim for all three categories – Power Seizers and the Constitutional Power Successors, Nonconstitutional Power Successors – to capture regimes with similar range of XCONST excluding the particular regime category. The xconst-sim variable is used in statistical analyses in Chapter 5 to contrast the effects of Power Seizers and Power Successors on international conflicts compared to other regimes with similar executive constraints. In sum, to the extent that the rates of conflict initiation differ between the Power Seizers and the Power Successors, we can attribute the difference to the personal traits of the Power Successors rather than to the institutional constraints on the leader. 38 Figure 4: Change of XCONST Over Time in the Power Successors (Constitutional) Compared to Democracies and Other Non-Democracies 39 Figure 4 further offers a glimpse into the institutional constraints in traditional monarchies compared to other non-democracies, 31 and democracies using the XCONST score of the Polity IV Dataset (Marshall, Jagger and Gurr 2009). 32 .It is to be noted that executive power remains unconstrained in most of the regimes over time with the exception of Egypt. Figure 5 provides another descriptive way to explore the extent of domes- tic constraints faced by autocrats. The figure illustrates the change of XCONST over time across the Power Seizers and the Power Successors in the nine “nonconstitutional” hereditary regimes. 33 These nine nonconstitutional hereditary regimes offers the merit of observing direct power transition from father to son in a given country controlling for a number of external factors. The vertical line marks the year in which power succession took place from father to son in each country. The XCONST score remains constant in six out of the nine countries, confirming that any changes in foreign policy paths can- not be owed to institutional arrangements in these countries. Taiwan and Togo saw one – while in Syria two – notch going upward in the XCONST score 34 after the second generation claimed power. 31 Regimes with polity2 score lower than 6. 32 XCONST is a seven-category scale with 1 meaning unlimited authority and 7 refer- ring to executive parity 33 These nine countries are the "nonconstitutional" Power Successors and their pre- ceding Power Seizers. 34 XCONST score 1 refers to “unlimited authority”; 2 “intermediate category”; and 3 “slight to moderate limitation on executive authority”. 40 Figure 5: Change of XCONST in the Nonconstitutional Dynasties Across the Hered- itary Power Seizers and the Power Successors (Nonconstitutional) Over Time) 41 Operationalizing Main IVs: Power Seizers and Power Successors Power Seizers Using Colgan’s Revolutionary Leader Dataset (2012), I code the Power Seizers whose main political quality is to rise through the ranks by means of force, such as coups, as- sassinations, and revolts motivated by personal ambition and dispositions such as risk tolerance. Colgan’s dataset is useful for my theory in that it is designed to identify revolutionary leaders and governments without reference to their foreign policy behav- ior (Colgan 2012, 452), reducing the concern for reverse causation between my IV (leader) and DV (international conflict). Moreover, the dataset contains several vari- ables that are useful to identify the Power Seizers. 35 I code leaders as Power Seizers if they came to power through irregular transition by forceful means, but not through popular demonstrations (IrregularTransition=1 & Usedforce=1). 36 Unlike Colgan’s rev- 35 For example, Colgan codes a leader as revolutionary (RevolutionaryLeader=1) 1) if the leader came to power in an irregular transition such as through use of armed force (Usedforce=1) or widespread popular demonstrations (IrregularTransition=1), and 2) if the leader implemented radical policy changes (RadicalPolicy=1), and 3) if the leader was not installed by a foreign power (ForeignInstalled=1), and 4) if the leader is not the founding leader of the state (FoundingLeader=0). These coding criteria are neces- sary for Colgan’s conceptualization of revolutionary leader who he defines as one that “transforms the existing social, and economic relationships of the state by overthrowing or rejecting the principal existing institutions of society” (Colgan 2012, 446) involving state-religion relationship, property ownership, and political ideology, etc. 36 Among the Power Seizers, Lee Kuan Yew of Singapore, Heydar Aliev of Azerbai- jan, and Francois Duvalier of Haiti are not coded as “Usedforce=1” in Colgan’s dataset. Nevertheless, I include them in the Power Seizers because of their well-known dra- matic rise to power as well as the supremacy over domestic political systems during their reigns. The result doesn’t change substantively if I drop these three first generation autocrats from the sample. 42 olutionary leader, however, leaders qualify as the Power Seizers in my theory even in the absence of sweeping domestic reforms implemented to transform social, economic, and political orders. Thus, autocrats such as Saddam Hussein, Idi Amin, and Muam- mar Qaddafi are coded both as revolutionary leaders in Colgan’s dataset and as Power Seizers in my data. Others who are repressive but are not necessarily known as hav- ing overthrown existing political order other than political rhetoric, such as Indonesia’s Suharto and Pakistan’s Musharraf, are not coded as revolutionary leader in Colgan’s, but are coded as Power Seizers in my dataset. All in all, the category for the Power Seizers is broader than that of the revolutionary leaders because of the omitted criteria of radical policy changes, and is almost inclusive of all of Colgan’s revolutionary leaders. 37 Constitutional Power Successors I use the Polity IV Dataset (Marshall, Jagger and Gurr 2009) to identify the Power Suc- cessors. Other prominent sources where the Power Successors are coded, commonly under the label of “monarchy”, include Geddes, Wright, Frantz’ Autocratic Breakdown and Regime Dataset (2014, GWF hereafter), Cheibub, Gandhi, and Vreeland’s Democ- racy and Dictatorship Dataset (2010, CGV hereafter), and Hadenius, Teorell, and Wah- man’s Authoritarian Regime Dataset (2013, HTW hereafter). The full list of monarchs among these three noted datasets differ from each other because of the varying cri- teria the authors used to classify monarchs. 38 I refrain from using these classification 37 Power Seizers consist of 2039 leader-year observations (1946-2004), whereas rev- olutionary leaders are 995 leader-year observations. For the full list of Power Seizers, see Appendix Table 14 38 CGV , for example, conceptualize regime types in terms of the “inner sanctums where real decisions are made and potential rivals are kept under close scrutiny” (Cheibub, Gandhi and Vreeland 2010, 84) for which monarchs, they determined, rely 43 criteria because of their broader concept employed for monarchs than what my theory calls for. All three datasets have additional layer of dimensions for monarchical regime, such as methods of power control (2014; 2013), or the methods of government removal (Cheibub, Gandhi and Vreeland 2010), other than the mode of power acquisition, which is the key criteria for distinguishing Power Successors. Instead, the Polity Dataset allows classification of monarchs whose criterion is solely based on the selection of the executives. I build on Weeks’ (2008, 61) coding practice of monarchies and draw on three component variables in the Polity Dataset – XRREG (Regulation of Chief Executive Recruitment); XRCOMP (Competitiveness of Executive Recruitment); and XROPEN (Openness of Executive Recruitment). Power Successors are coded as 1 if the chief executive assumes their position of power by right of descent or if the executive recruitment is determined by hereditary succession and designation. 39 This coding practice excludes all monarchs that are head of state, but not head of government, as well as monarchies where the balance of power within the executive branch favors the chief minister over the monarch. The full list of Power Successors (1946-2010) include leaders from 18 countries, many of them in the Middle on family and kin networks to acquire power and maintain it. GWF define regime as “basic informal and formal rules that determine what interests are represented” (Ged- des, Wright and Frantz 2014, 314) and group autocratic regime cases as either party, military, personalist, or monarchical. HTW, on the other hand, focus on the “modes of accessing and maintaining political power” (Wahman, Teorell and Hadenius 2013, 20), and classify monarchies based on their hereditary power succession along with military regimes and electoral regimes. 39 1) XRREG=3(Regulated) & XRCOMP=1(Selection) & XROPEN=1 (Closed), or 2) XRREG=3(Regulated) & XRCOMP=1(Selection) & XROPEN=2(Dual Executive- Designation). Additionally, I leave out Kim Jong-il and Kim Jung-un of North Korea and Jean-Claude Duvalier of Haiti from the return of this particular coding that qualify as the nonconstitutional Power Successors. 44 East, but also countries from Asia or Africa, such as Nepal and Swaziland (Table 15 in Appendix). 40 Nonconstitutional Power Successors Drawing on Brownlee’s data (Brownlee 2007) that contains 258 post-WWII autocrats, I collect a total of 11 ”nonconstitutional” Power Successors from 9 countries with Nicaragua and North Korea having inaugurated a third generation of hereditary descent (Table 2). Nonconstitutional hereditary successions are coded as such when they entail de facto power transfer even when a figurehead placeholder was installed as a cover. Lee Kuan Yew of Singapore, for example, retained elder statesman post and continued to overshadow the formal premiership of Goh Chok Tong who was in the post until 2004 and gave way to Lee Hsein Loong, the eldest son of Lee Kuan Yew. Hereditary succes- sion is confined to father-son leadership transition that was deliberately implemented by the ruler before his death. Accordingly, a fraternal succession, the only case of which is Raul Castro of Cuba to succeed Fidel Castro is omitted from the dataset. 41 Lastly, since the deliberate initiation of power transfer in the absence of formal election is the central element to hereditary succession, persistent influence of political dynasties in electoral democracies such as the Philippines, South Korea, and India are omitted. 42 40 Among the 18 countries, 15 countries overlap with GWF classification of monar- chies (1946-2010); all 18 countries are included in CGV’s royals classification (1946- 2008), and 13 countries in the HTW dataset (1972-2010). For the full list of Power Successors, see Appendix Table 15. 41 The result doesn’t change substantively even when the case is included in the anal- ysis. 42 Corazon Aquino (1986-1992) and Benigno Aquino III (2010-present) in the Philip- pines; Park Jung-hee (1962-1979) and Park Geun-hye (2013-present) in South Korea; 45 Dependent Variable and Model Specification As my theoretical discussion is monadic in that it focuses on the initiating country’s leadership type rather than the interactions of different polities in a dyad, I test my argument using a country-leader-year set-up. For data on military conflict, I use the Militarized Interstate Disputes (MID) dataset (Palmer et al. 2015), which records all instances of when one state threatened, displayed, or used force against another. My primary dependent variable counts the number of MIDs that a country initiated in a given year. I also test the hypotheses using higher-levels of conflict as DV , which counts MIDs that involve actual display or use of force and war only. To control for temporal interdependence, I include cubic splines of the number of years since the last time state A initiated a MID (Beck, Katz and Tucker 1998; Carter and Signorino 2010) and clus- ter the standard errors by country to account for unobserved differences across states. For every hypothesis test, I carry out the analysis with fixed effects to correct for the possibility of omitted variables specific to the initiating country. Lastly, I use a negative binomial model because the dependent variables are event counts (the number of MIDs) and the negative binomial model accounts for overdispersion in my dependent variables (Scott Long 1997). Robustness Check Alternative Explanation: Domestic Stability A skeptical reader may well point out that the domestic and international instability cre- ated by a newly established regime could affect the regime’s propensity for international conflict, rather than the leader’s predisposition towards use of force itself. If this is the Jawaharlal Nehru (1947-1964) and Indira Gandhi (1980-1984) in India. 46 case, there will be more international conflicts under the Power Seizers because of the unsettling condition fueled by the forceful and violent process of attaining power. As time goes by, however, domestic conditions will eventually settle down, which leads to less international conflicts by the time the Power Successors assume power. This alter- native explanation merits attention since newly established regimes can indeed be more unstable domestically, and such domestic turmoil may spill over into its international relations (Colgan 2010) as part of an attempt to, for example, divert the public’s and opponents’ attention from domestic political troubles (Levy and Vakili 2014). The core of this alternative explanation concerns the notion of how much time has passed since the establishment of a new regime. Thus, I include a variable called Regime Stability in all of my analyses that counts the years since the establishment of a regime by the Power Seizers Power Successors to account for the potential addition of insecurity and conflict-proneness of a young regime. Reverse Causation It may be possible that the already frequently existing international conflicts surrounding a regime prompt ambitious individuals to challenge domestic status quo and seize power amidst turmoil, rather than the other way around. Moreover, it seems plausible that the rise of Power Seizers is affected by regional dynamics where certain geographic regions with a history of rivalry and weak institutions lead more ambitious individuals to emerge. In order to mitigate these possible concerns of endogeneity, I include regional dummy variables and lag all of my predictor variables with the exception of the “Regime Stability” variable that counts the passage of years since regime establishment. 47 Standard Control Variables Following the bulk of the IR literature, 43 I include a number of control variables that could affect a state’s decision to begin military hostilities. In an attempt to address the concern for endogeneity, I lag all of the control variables by one year except for regime stability which counts the years since the establishment of the regime. Capabilities Powerful states are more likely to initiate conflict than others (Most and Starr, 1989). To control for the variation in power capabilities, I include the raw military capabilities score of the country using the Correlates of War (COW) Composite Index of National Capabilities (CINC) data. This score uses six measures of a state’s power and divides it by the total measure of the world. For example, a state’s share of military personnel is its total number of military personnel divided by the total number of military personnel worldwide. Major Power Status Research has shown that powerful states have wider ranging interests and thus are more prone to conflict than minor powers. Relying on the COW State System Membership data, a dummy variable indicating whether the state is a major power is included to control for a state’s major power status. 43 See Bennett and George (2005); Lai and Slater (2006); Gleditsch, Salehyan and Schultz (2008); Gowa (2000); Oneal, Russett and Berbaum (2003); Gleditsch (2002); Colgan and Weeks (2015) 48 Alliances Scholars have argued that having more allies is likely to provide more opportunities for conflict. Allies may embolden countries to engage in more conflict with the belief that their allies would back them up in case of military disputes (Leeds 2003; Snyder Glenn 1997). The measure of alliance may also be proxy for geopolitical interests (Gowa 2000). Using COW Formal Alliances data, the alliances variable counts the country’s number of alliance partners. Contiguity Previous research has pointed to contiguity as an important indicator of the likelihood a state will engage in conflict (Huth 1996). Thus, I control the number of contiguous borders, which is the total number of states by land or sea, across no more than 400 miles of water, as defined by the COW project. Openness to Trade A long tradition in IR theory indicates that trade inhibits conflicts among states. Trade dependent states should be less likely to get involved in conflicts to preserve the benefits of trade for their economies (Oneal and Russett 2001). To control for the trade depen- dency, a measure of economic openness is included, which is a state’s total trade (exports + imports) divided by its GDP. Data on countries’ exports and imports are drawn from COW project, while GDP data is from the Gleditsch GDP data. 44 44 Exports and imports data from COW covers from 1870 to 2009; while Gleditsch GDP data range from 1950-2011. As a result, the data are missing for years from 1945- 1949 and 2010. 49 Regime Stability Newly established regimes tend to be unstable domestically, and such domestic turmoil may spill over into its international relations (Colgan 2012) as part of an attempt to, for example, divert the public’s and opponents’ attention from domestic political troubles (Levy and Vakili 2014). A variable that counts the years since the establishment of a regime is included to account for the potential addition of insecurity and conflict- proneness of a young regime. Region Dummies Regional dummy variables are included as part of an attempt to mitigate the concern for endogeneity (Colgan and Weeks 2015) – that international conflicts cause leadership to change domestically. For example, leaders from a geographic region that is prone to international conflicts might be subject to political turmoil domestically as well, and thus are more likely to be replaced leading to frequent occurrences of first generation leaders. To control for this geographic effect, I insert regional dummy variables. 50 Chapter 5: Empirical Results H1: Power Seizers vs. Power Successors: The Power Seizers are more likely to initiate international conflicts than the Power Succes- sors. The first hypothesis to assess is that the Power Seizers are more likely to initiate in- ternational conflicts than the Power Successors (H1). First, I compare the coefficients of Power Seizers with those of Power Successors (Constitutional) / (Nonconstitutional) through two-tailed Wald Tests, and evaluate the significance of the differences in these coefficients (Table 2). Next, I exponentiate the coefficients to interpret the effects of predictor variables in negative binomial regressions by recovering the difference in the log of expected counts of MIDs for a one unit change. 45 For example, the coefficient of 0.315 for Power Seizers in model (1) (Table 3) indicates that the rate for MID would be expected to increase by a factor of 1.37, 46 or 37% of rate, if the leader is a Power Seizer, while holding all other variables constant. I predict that the difference of coefficient estimates between Power Seizers and Power Successors (Constitutional) / (Nonconstitutional) is statistically significant through Wald Tests with the overall estimates of the Power Seizers greater than those of the Power Successors. The results in Table 3 are robust both statistically and substan- tively confirming my first hypothesis (H1). The Wald Tests demonstrate a clear statisti- cal difference in MID initiation by Power Seizers and by Power Successors(Constitutional) at p<0.0000 across all models including the fixed effects analyses. Substantively, the dif- 45 The coefficients in negative binomial regressions are usually reported in terms of incidence rate ratios, which can be written as b = log(m x 0 +1 ) log(m x 0 ) = log(m x 0 +1 /m x 0 ), where b is the regression coefficient,m is the expected count. 46 The exponent of 0.315 is 1.37. 51 ference between Power Seizers and Power Successors (Constitutional) is as stark. Com- pared to Power Successors (Constitutional), Power Seizers are more likely to initiate MIDs by as high as a whopping 252% increase in rate and by as low as 159%. 47 . For the ”nonconstitutional” Power Successors, the coefficient estimates for the Power Seizers are greater than Power Successors (Nonconstitutional) across all models as expected, marking as high as 81% increase and as low as 33.5% increase in rate of MID initiation by Power Seizers compared to Power Successors (Nonconstitutional). 48 However, Wald Tests fail to confirm statistical significance in the difference between Power Seizers and Power Successors (Nonconstitutional) in models (5), (6), and (7), due to the larger standard errors pointing to the smaller number of observations from nine countries only (see Table 2). 49 Figure 6 simulates the substantive effects using CLARIFY (Tomz et al. 2003). I estimate the expected probability of MID initiation when all of the covariates are set to the observed values for Iraq in 1990. 50 The results indicate that, had Iraq been ruled by a ”constitutional” Power Successor in 1990 instead of Saddam Hussein, who was a Power Seizer, there would have been a 27.47 % decrease in chance (from 61% to 33.53%) that 47 The % change is calculated using the formula:(e bb 0 1)100, where b is the coefficient estimate of Power Seizers and b’ is the coefficient estimate of Power Succes- sors (Constitutional). The calculated % change is 161% for model (1); 252% increase for model (2); 159% for model (3); and 243% for model (4). 48 An increased rate of 33.5% for model (5); 59.3% for model (6); 47.4% for model (7); 81% for model (8). 49 2152 state-year observations for Power Seizers; 101 observations for Power Succes- sors (Nonconstitutional); and 657 observations for Power Successors (Constitutional). 50 The figures are based off of model (1) and model (4) in Table 3. 52 Table 3: MID Initiation by Power Seizers vs. Power Successors (1) (2) FE (3) (4) FE (5) (6) FE (7) (8) FE Any MIDs Any MIDs Higher Levels of MIDs Higher Levels of MIDs Any MIDs Any MIDs Higher Levels of MIDs Higher Levels of MIDs Power Seizers 0.315 a 0.207 b 0.345 c 0.241 d 0.394 e 0.268 f 0.419 g 0.296 h (0.085) (0.071) (0.086) (0.073) (0.093) (0.072) (0.093) (0.073) Power Successors (Constitutional) -0.644 a -1.050 b -0.607 c -0.990 d (0.166) (0.231) (0.175) (0.234) Power Successors (Nonconstitutional) 0.105 e -0.197 f 0.031 g -0.297 h (0.289) (0.276) (0.227) (0.292) Trade Openness -0.353 -0.119 -0.365 -0.075 -0.336 -0.083 -0.347 -0.030 (0.152) (0.095) (0.162) (0.107) (0.144) (0.086) (0.152) (0.098) Years since Regime Est. 0.002 0.004 t 0.003 0.004 t 0.001 0.003 0.001 0.003 (0.003) (0.002) (0.003) (0.002) (0.003) (0.002) (0.003) (0.002) Capability 2.487 -0.276 2.504 -0.138 2.218 t -0.285 2.244 t -0.166 (1.116) (1.262) (1.112) (1.291) (1.177) (1.264) (1.168) (1.293) Total Borders 0.067 0.065 0.074 0.073 (0.016) (0.017) (0.017) (0.017) Major Power Status 0.422 19.839 0.432 13.731 0.395 22.689 0.406 14.691 (0.130) (.) (0.128) (389.029) (0.152) (.) (0.148) (851.822) Total Allies 0.010 0.005 0.013 0.011 0.009 -0.000 0.012 0.006 (0.017) (0.016) (0.017) (0.016) (0.018) (0.016) (0.018) (0.016) Years Since Last MID -0.022 0.042 -0.021 0.043 -0.022 0.041 -0.022 0.042 (0.010) (0.013) (0.010) (0.013) (0.010) (0.013) (0.010) (0.013) Cubic Splines YES YES YES YES YES YES YES YES Region Dummies YES NO YES NO YES NO YES NO Constant -1.195 1.306 -1.247 1.276 -1.244 1.103 -1.291 1.085 (0.143) (0.209) (0.143) (0.219) (0.147) (0.190) (0.147) (0.199) Constant -0.957 -0.931 -0.902 -0.880 (0.343) (0.346) (0.341) (0.345) Observations 7540 6922 7540 6874 7540 6922 7540 6874 Negative binomial regressions with standard errors in parentheses clustered by country. t p< 0.10, p< 0.05, p< .01, p< .001 Models (2), (4), (6), and (8) are fixed effects models of (1), (3), (5), and (7), and thus drop time-invariant variables such as contiguity (Total Borders) and region dummies. a Wald Test: different at p< 0.0000. b Wald Test: different at p< 0.0000. c Wald Test: different at p< 0.0000. d Wald Test: different at p< 0.0000. e Wald Test: different at p< 0.3073. f Wald Test: different at p< 0.0872. g Wald Test: different at p< 0.0789. h Wald Test: different at p< 0.0396. 53 Iraq would invade Kuwait despite similar institutional constraints. The point estimates for the ”nonconstitutional” Power Successors are also in the same expected directions, although as discussed earlier, they fall short of demonstrating statistical significance due to larger confidence intervals around the estimates. The major takeaway is that despite similar institutional constraints, different leadership types have different effects on a country’s propensity for initiation of inter- national conflicts. In other words, Iraqi invasion of Kuwait in 1990 can be attributed as much to the fact that Saddam Hussein was a Power Seizer as to the fact that he was an unconstrained strongman as suggested by the existing literature (Weeks 2012). H3a: "Nonconstitutional" Power Successors vs. "Constitutional" Power Successors: The ”nonconstitutional” Power Successors are more likely to initiate international conflicts than the ”constitutional” Power Successors. Figure 6 reveals another interesting aspect about the nonconstitutional Power Successors. While the point estimates of the Power Seizers are very similar across both models (61 vs. 57.40), that of the Power Successors is much bigger for the “nonconstitu- tional” compared to the “constitutional” (48.48 vs. 33.53). This supports my hypothesis (H3a) that the “nonconstitutional” Power Successors are more belligerent than the “con- stitutional” Power Successors, albeit with greater variation. Finally, how do these different leadership types compare to democratic lead- ers? I conduct an analysis including all the predictor dummy variables with the base category set as “Democracy.” I hypothesize that the Power Seizers are more likely to initiate international conflicts (H2), while the Power Successors are no more likely to 54 Figure 6: Expected Percentage of Conflict Initiation: Iraq 1990 Scenario initiate conflicts than democratic leaders (H3). I further argue that among the Power Successors, the “nonconstitutional” successors are more belligerent than the “constitu- tional” successors (H3a). H2: Power Seizers vs. Democratic Leaders: The Power Seizers are more likely to initiate international conflicts than democratic leaders. The results from Table 4 support my hypotheses both statistically and substan- tively. Power Seizers are more likely to initiate international conflicts than democratic leaders by as high as 49.6% in rate and by as low as 28.3% in rate, 51 and the results are statistically significant (H2). 51 45.4% (model 9); 28.3% (model 10); 49.6% (model 11); 33.1% (model 12) 55 H3: Power Successors vs. Democratic Leaders: The Power Succes- sors are no more likely to initiate international conflicts than demo- cratic leaders. The results for the Power Successors from Table 4 also show consistency with H3. Power Successors (Constitutional) are less likely to initiate international conflicts than democracies by as high as 64% in rate, and by as low as 43.4% in rate, 52 and the results are highly statistically significant at either p<0.01 or p<0.001. Power Successors (Nonconstitutional), on the other hand, are statistically indistinguishable from democra- cies. Overall, evidence from Table 4 supports my hypothesis that the succeeding Power Successors are no more likely to initiate conflicts than democratic leaders (H3). Figure 7 illustrates the substantive effects based off of Model (9) from Table 4 using CLARIFY (Tomz et al. 2003). As in the earlier simulation, I set the control variables to the values for Iraq in 1990. The point estimates in Figure 7 show consis- tency with my hypotheses H2, H3, and H3a. The Power Seizers demonstrate a higher rate of conflict initiation than democracies confirming H2. However, the estimates differ for the Power Successors depending on whether they are ”constitutional” or ”noncon- stitutional” successors. Note that the Power Successors (Constitutional) are the most peaceful by a large margin, even compared to democracies. However, the point esti- mates for the Power Successors (Nonconstitutional) are statistically indistinguishable from the democracies and even from the Power Seizers. This pattern confirms H3a that the nonconstitutional Power Successors are more belligerent than the constitutional Power Successors. In other words, although Kim Jung-un of North Korea may be less likely to initiate international disputes compared to his grandfather, Kim Il-sung who 52 45.4% (model 9); 64% (model 10); 43.4% (model 11); 61.6 % (model 12) 56 Table 4: Power Seizers and Power Successors vs. Democracies (9) (10) FE (11) (12) FE Any MIDs Any MIDs Higher Levels of MIDs Higher Levels of MIDs Power Seizers 0.374 0.249 0.403 0.286 (0.107) (0.100) (0.108) (0.103) Power Successors (Constitutional) -0.606 -1.023 -0.570 -0.956 (0.171) (0.245) (0.182) (0.249) Power Successors (Nonconstitutional) 0.111 -0.229 0.041 -0.319 (0.309) (0.283) (0.249) (0.299) Other Autocrats 0.112 0.111 0.114 0.119 (0.107) (0.099) (0.105) (0.102) Trade Openness -0.346 -0.108 -0.356 -0.059 (0.151) (0.092) (0.161) (0.104) Years since Regime Est. 0.003 0.004 t 0.004 0.004 t (0.003) (0.002) (0.003) (0.002) Capability 2.402 -0.484 2.408 -0.366 (1.165) (1.270) (1.170) (1.301) Total Borders 0.065 0.064 (0.016) (0.017) Major Power Status 0.430 14.528 0.440 24.696 (0.141) (584.518) (0.138) (.) Total Allies 0.008 0.006 0.011 0.012 (0.017) (0.016) (0.017) (0.016) Years Since Last MID -0.022 0.042 -0.022 0.042 (0.010) (0.013) (0.010) (0.013) Cubic Splines YES YES YES YES Region Dummies YES NO YES NO Constant -1.256 1.268 -1.305 1.232 (0.157) (0.222) (0.157) (0.232) Constant -0.955 -0.930 (0.342) (0.345) Observations 7490 6884 7490 6836 Negative binomial regressions with standard errors in parentheses clustered by country. t p< 0.10, p< 0.05, p< .01, p< .001 Models (2) and (4) are fixed effects of models of (1) and (3), and thus drop time-invariant variables such as contiguity and region dummies. 57 was a Power Seizer, he might be more willing to engage in international conflicts com- pared to his counterparts in the constitutional monarchies. Figure 7: Expected Percentage of Conflict Initiation Compared to Democracy: Iraq 1990 Scenario H4: Power Seizers vs. All Other Autocrats Under Similar Constraints: The Power Seizers are more likely to initiate international conflicts than all other autocrats facing similar institutional constraints. In order to assess H4, I create a variable Similar Constraints that captures autocrats with XCONST value smaller or equal to 3 53 excluding all the Power Seizers. 53 Note that the smaller the XCONST value, the higher the institutional constraints on the executive. 58 I compare the coefficients of Power Seizers with those of Similar Constraints, carrying out two-tailed Wald Tests to evaluate the significance of the differences in these coefficients (Table 5). Next, I exponentiate the coefficients to interpret the effects of pre- dictor variables in negative binomial regressions by recovering the difference in the logs of expected counts of MIDs for one unit change. 54 For example, the coefficient of 0.323 for Power Seizers in model (1) (Table 5) indicates that the rate for MID would be ex- pected to increase by a factor of 1.381, 55 or 38.1% of rate, if the leader is a Power Seizer, while holding all other variables constant. I expect that the coefficient estimates for Power Seizers to be larger than those of Similar Constraints, and more importantly, that the differences between these coefficients to prove statistically significant through Wald Test. I do not expect, however, that the coefficient estimates for Similar Constraints to be necessarily negative, nor those for Power Seizers to be positive. The results from Table 5 indicate that MIDs are more likely to occur by as low as 31.5% and as high as 43.9% 56 in rate if the Power Seizers are in power com- pared to when they are not. These effects are highly statistically significant across all models. Furthermore, the Wald Tests indicate that the differences between the Power Seizers and autocrats under similar institutional constraints are statistically significant at either p<0.0000 or p<0.0005. Substantively, the differences are as striking. There is an increased rate that MIDs will occur by Power Seizers than by Similar Constraints 54 The coefficients in negative binomial regressions are usually reported in terms of incidence rate ratios, which can be written as b = log(m x 0 +1 ) log(m x 0 ) = log(m x 0 +1 /m x 0 ), where b is the regression coefficient,m is the expected count. 55 The exponent of 0.323 is 1.381. 56 38.1% for model (1) ; 31.5% for model (2); 43.9% for model (3); and 37.8% for model (4). 59 by a percentage rate ranging from 32.5% to 60.8%. 57 Thus, the evidence supports my hypothesis that the Power Seizers are more likely to initiate international conflicts than other autocrats under similar institutional constraints. 57 The % change is calculated using the formula:(e bb 0 1)100, where b is the coefficient estimate of Power Seizers and b’ is the coefficient estimate of Similar Con- straints. An increased rate of 59% for model (1); 32.5% for model (2); 60.8% for model (3); and 35.4% for model (4). 60 Table 5: MID Initiation by Power Seizers vs. Autocrats Under Similar Institutional Constraints (1) (2) (3) (4) Any MIDs Any MIDs Higher Levels of MIDs Higher Levels of MIDs Power Seizers 0.323 a 0.273 b 0.364 c 0.320 d (0.105) (0.086) (0.107) (0.089) Similar Constraints -0.141 a -0.008 b -0.111 c 0.017 d (0.106) (0.090) (0.106) (0.093) Trade Openness -0.349 -0.085 -0.358 -0.033 (0.143) (0.086) (0.152) (0.099) Years since Regime Est. 0.001 0.003 0.001 0.004 (0.003) (0.002) (0.003) (0.002) Capability 2.481 -0.248 2.460 -0.172 (1.154) (1.277) (1.151) (1.308) Total Borders 0.073 0.072 (0.017) (0.017) Major Power Status 0.395 14.532 0.406 15.060 (0.136) (823.857) (0.136) (1466.101) Total Allies 0.012 -0.001 0.014 0.006 (0.017) (0.016) (0.017) (0.016) Years Since Last MID -0.022 0.042 -0.022 0.043 (0.010) (0.013) (0.010) (0.013) Cubic Splines YES YES YES YES Region Dummies YES NO YES NO Constant -1.176 1.099 -1.239 1.062 (0.150) (0.196) (0.151) (0.205) lnalpha Constant -0.914 -0.889 (0.340) (0.344) Observations 7540 6922 7540 6874 Negative binomial regressions with standard errors in parentheses clustered by country. t p< 0.10, p< 0.05, p< .01, p< .001 Models (2), (4) are fixed effects models of (1), (3) respectively, and thus drop time-invariant variables such as contiguity (Total Borders) and region dummies. a Wald Test: different at p< 0.0000. b Wald Test: different at p< 0.0005. c Wald Test: different at p< 0.0000. d Wald Test: different at p< 0.0002. 61 H5: Power Successors vs. All Other Autocrats Under Similar Con- straints: The Power Successors are less likely to initiate interna- tional conflicts than all other autocrats facing similar institutional constraints. The next hypothesis is that the Power Successors (Constitutional) / (Noncon- stitutional) - are less likely to initiate conflict than other autocrats facing similar institu- tional constraints on the executives - Similar Constraints (Constitutional) / (Nonconsti- tutional). For this hypothesis testing, I create two variables called Similar Constraints (Constitutional) / (Nonconstitutional) each of which captures regimes with XCONST value smaller or equal to 3 58 excluding either the constitutional Power Successors or the nonconstitutional ones. Note that the mean XCONST value for the constitutional Power Successors is 1.80 with standard deviation at 1.07; and the mean at 1.77 and standard deviation at 0.91 for the nonconstitutional Power Successors. I predict that the coefficient estimates for Power Successors (Constitutional) / (Nonconstitutional) are smaller than those of Similar Constraints (Constitutional) / (Nonconstitutional), and that their differences are statistically significant through Wald Test. The results from Table 6 strongly support my hypothesis both statistically and substantively. Power Successors (Constitutional) are less likely to initiate MIDs com- pared to Similar Constraints (Constitutional) by more than 58% in rate across all mod- els, and the rates for the fixed effects models are even stronger indicating a decrease of 71.3% in rate for model (2) and 70.5% for model (4). 59 Wald Tests indicate that these 58 Note that the smaller the XCONST value, the higher the institutional constraints on the executive. 59 A decrease of 58.2% in rate for model (1); 71.3% for model (2); 58% for model (3); 70.5% for model (4). 62 differences are statistically significant at p<0.0000 for all four models. The coefficient estimates of the nonconstitutional Power Successors are also in line with my prediction suggesting a decrease in rate ranging from 9.9% to 42.2% across models (5) to (8). 60 However, the differences fall short of achieving statistical significance in Wald Tests. 61 60 A decrease of 9.9% in rate for model (5); 34.3% for model (6); 18.6% for model (7); 42.2% for model (8). 61 101 observations for Power Successors (Nonconstitutional), 657 for Power Succes- sors (Constitutional). 63 Table 6: MID Initiation by Power Successors vs. Autocrats under Similar Institu- tional Constraints (1) (2) FE (3) (4) FE (5) (6) FE (7) (8) FE Any MIDs Any MIDs Higher Levels of MIDs Higher Levels of MIDs Any MIDs Any MIDs Higher Levels of MIDs Higher Levels of MIDs Successors -0.689 a -1.039 b -0.643 c -0.975 d (Constitutional) (0.163) (0.232) (0.172) (0.235) Similar Constraints 0.181 a 0.208 b 0.223 c 0.245 d (Constitutional) (0.083) (0.076) (0.082) (0.078) Successors -0.007 e -0.267 f -0.066 g -0.355 h (Nonconstitutional) (0.282) (0.278) (0.221) (0.294) Similar Constraints 0.096 e 0.152 f 0.139 g 0.193 h (Nonconstitutional) (0.086) (0.075) (0.087) (0.078) Trade Openness -0.358 -0.125 -0.367 -0.084 -0.357 -0.086 -0.363 -0.032 (0.151) (0.096) (0.161) (0.109) (0.149) (0.086) (0.155) (0.099) Years since Regime Est. 0.002 0.005 0.003 0.005 -0.001 0.004 t -0.000 0.004 t (0.003) (0.002) (0.003) (0.002) (0.003) (0.002) (0.003) (0.002) Capability 1.861 -0.790 1.790 -0.745 1.392 -0.770 1.308 -0.748 (1.171) (1.272) (1.201) (1.303) (1.182) (1.268) (1.212) (1.299) Total Borders 0.066 0.065 0.077 0.076 (0.016) (0.017) (0.018) (0.018) Major Power Status 0.456 14.108 0.467 14.856 0.432 13.825 0.442 13.833 (0.131) (503.577) (0.132) (773.758) (0.143) (547.442) (0.145) (463.097) Total Allies 0.006 0.006 0.009 0.012 0.006 -0.001 0.008 0.006 (0.018) (0.016) (0.018) (0.016) (0.019) (0.016) (0.019) (0.016) Years Since Last MID -0.022 0.044 -0.021 0.044 -0.024 0.042 -0.023 0.042 (0.010) (0.013) (0.010) (0.013) (0.010) (0.013) (0.010) (0.013) Cubic Splines YES YES YES YES YES YES YES YES Region Dummies YES NO YES NO YES NO YES NO Constant -1.147 1.261 -1.210 1.218 -1.087 1.077 -1.147 1.037 (0.137) (0.210) (0.137) (0.219) (0.144) (0.187) (0.144) (0.196) Constant -0.917 -0.890 -0.813 -0.787 (0.348) (0.351) (0.355) (0.359) Observations 7542 6924 7542 6876 7542 6924 7542 6876 Negative binomial regressions with standard errors in parentheses clustered by country. t p< 0.10, p< 0.05, p< .01, p< .001 Models (2), (4), (6), and (8) are fixed effects models of (1), (3), (5), and (7), respectively, and thus drop time-invariant variables such as contiguity (Total Borders) and region dummies. a Wald Test: different at p< 0.0000. b Wald Test: different at p< 0.0000. c Wald Test: different at p< 0.0000. d Wald Test: different at p< 0.0000. e Wald Test: different at p< 0.7074. f Wald Test: different at p< 0.1242. g Wald Test: different at p< 0.3355. h Wald Test: different at p< 0.0578. 64 H6:Distinguishing Power Seizers: The Power Successors in noncon- stitutional dynasties are less likely to initiate international conflicts than their immediate preceding Power Seizers. Although not specified in my theoretical framework (Chapter 3), one could further distinguish among the Power Seizers based on whether their descendants are the “nonconstitutional” Power Successors (Table 2). I call those Power Seizers, such as Kim Il-Sung, Hafez al-Assad, and Francois Dvualier, whose descendants are “nonconstitu- tional” successors (see Table 2) the “hereditary” Power Seizers, and most other Power Seizers who have failed to pass down power to their descendants the “nonhereditary” Power Seizers. The main purpose of this distinction is to explicitly collect Nonconsti- tutional Dynasties which are made up of the “hereditary” Power Seizers and the “non- constitutional Power Successors (Table 2). This distinction offers the merit of exploring any differences in the conflict propensity by the first generation, hereditary Power Seiz- ers and the succeeding nonconstitutional Power Successors in those Nonconstitutional Dynasties which happen to be the most belligerent regimes, such as North Korea. Within these nonconstitutional dynasties, I expect that the succeeding Power Successors are less conflict prone compared to the Power Seizers, their immediate predecessor. 62 First, I compare the coefficients of Power Seizers (Nonhereditary) / (Hered- itary) with those of Power Successors (Constitutional) / (Nonconstitutional) through two-tailed Wald Tests, and evaluate the significance of the differences in these coeffi- cients (Table 7). Next, I exponentiate the coefficients to interpret the effects of predictor variables in negative binomial regressions by recovering the difference in the log of ex- 62 I carry out additional hypothesis testing based on distinguished Power Seizers in the Appendix of this dissertation. For example, In the Appendix Table 12, I show that the "hereditary" and the "nonhereditary" Power Seizers are more likely to initiate conflict compared to other autocrats under similar constraints. 65 pected counts of MIDs for one unit change. 63 For example, the coefficient of 0.285 for Power Seizers (Nonhereditary) in model (1) (Table 7) indicates that the rate for MID would be expected to increase by a factor of 1.33, 64 or 33% of rate, if the leader is a nonhereditary Power Seizer, while holding all other variables constant. I predict that the difference of coefficient estimates between Power Seizers (Nonhereditary) / (Hereditary) and Power Successors (Constitutional) / (Nonconstitu- tional) is statistically significant through Wald Tests with the overall estimates of the first generations greater than those of the succeeding generations. The results in Table 7 are robust both statistically and substantively. The Wald Tests demonstrate a clear statistical difference in MID initiation by Power Seizers (Nonhereditary) and by Power Succes- sors (Constitutional) at p<0.0000 across all models including the fixed effects analyses. Substantively, the difference between Power Seizers (Nonhereditary) and Power Succes- sors (Constitutional) is as stark. Compared to Power Successors (Constitutional), Power Seizers (Nonhereditary) are more likely to initiate MIDs by as high as a whopping 261% increase in rate and by as low as 159%. 65 Note again that the Nonconstitutional Dynasties are made up of the “hered- itary” Power Seizers and the “nonconstitutional” Power Successors (Table 2). For the Nonconstitutional Dynasties, the coefficient estimates for the Power Seizers (Heredi- 63 The coefficients in negative binomial regressions are usually reported in terms of incidence rate ratios, which can be written as b = log(m x 0 +1 ) log(m x 0 ) = log(m x 0 +1 /m x 0 ), where b is the regression coefficient,m is the expected count. 64 The exponent of 0.285 is 1.33. 65 The % change is calculated using the formula:(e bb 0 1)100, where b is the coefficient estimate of Power Seizers (Nonhereditary) and b’ is the coefficient estimate of Power Successors (Constitutional). The calculated % change is 161% for model (1); 261% increase for model (2); 159% for model (3); and 251% for model (4). 66 tary) are greater than Power Successors (Nonconstitutional) across all models as ex- pected, marking as high as 60.7% increase and as low as 25.4% increase in rate of MID initiation by Power Seizers (Hereditary) compared to Power Successors (Nonconstitu- tional). 66 However, Wald Tests fail to confirm statistical significance in the difference between Power Seizers (Hereditary) and Power Successors (Nonconstitutional) due to the larger standard errors pointing to the smaller number of observations from nine coun- tries only (see Table 2). 67 66 An increased rate of 43.7% for model (5); 25.4% for model (6); 60.7% for model (7); 43.3% for model (8). 67 233 state-year observations for Power Seizers (Hereditary), 1919 observations for Power Seizers (Nonhereditary); 101 observations for Power Successors (Nonconstitu- tional); and 657 observations for Power Successors (Constitutional). 67 Table 7: MID Initiation by Power Seizers vs. Power Successors (1) (2) FE (3) (4) FE (5) (6) FE (7) (8) FE Any MIDs Any MIDs Higher Levels of MIDs Higher Levels of MIDs Any MIDs Any MIDs Higher Levels of MIDs Higher Levels of MIDs Power Seizers (Nonhereditary) 0.285 a 0.250 b 0.309 c 0.279 d (0.091) (0.074) (0.092) (0.076) Power Successors (Constitutional) -0.677 a -1.034 b -0.646 c -0.977 d (0.162) (0.232) (0.170) (0.235) Power Seizers (Hereditary) 0.318 te -0.347 f 0.346 tg -0.309 h (0.174) (0.223) (0.177) (0.227) Power Successors (Nonconstitutional) -0.045 e -0.574 tf -0.129 g -0.669 h (0.284) (0.297) (0.221) (0.312) Trade Openness -0.359 -0.121 -0.372 -0.078 -0.369 -0.102 -0.382 -0.058 (0.151) (0.095) (0.161) (0.108) (0.151) (0.090) (0.157) (0.102) Years since Regime Est. 0.002 0.004 t 0.003 0.004 t -0.001 0.004 -0.001 0.004 (0.003) (0.002) (0.003) (0.002) (0.003) (0.002) (0.003) (0.002) Capability 2.493 -0.202 2.509 -0.060 1.504 -0.461 1.478 -0.364 (1.098) (1.264) (1.095) (1.293) (1.122) (1.255) (1.114) (1.284) Total Borders 0.064 0.062 0.079 0.078 (0.016) (0.017) (0.017) (0.017) Major Power Status 0.425 14.626 0.435 14.485 0.442 13.676 0.455 13.505 (0.126) (836.192) (0.123) (534.961) (0.133) (454.544) (0.129) (435.625) Total Allies 0.012 0.005 0.015 0.011 0.006 -0.001 0.009 0.006 (0.016) (0.016) (0.016) (0.016) (0.018) (0.016) (0.019) (0.016) Years Since Last MID -0.022 0.042 -0.021 0.043 -0.024 0.041 -0.023 0.041 (0.010) (0.013) (0.010) (0.013) (0.010) (0.013) (0.010) (0.013) Cubic Splines YES YES YES YES YES YES YES YES Region Dummies YES NO YES NO YES NO YES NO Constant -1.145 1.304 -1.190 1.275 -1.063 1.188 -1.099 1.169 (0.153) (0.211) (0.154) (0.220) (0.129) (0.182) (0.128) (0.188) Constant -0.949 -0.921 -0.816 -0.789 (0.343) (0.346) (0.358) (0.363) Observations 7542 6924 7542 6876 7542 6924 7542 6876 Negative binomial regressions with standard errors in parentheses clustered by country. t p< 0.10, p< 0.05, p< .01, p< .001 Models (2), (4), (6), and (8) are fixed effects models of (1), (3), (5), and (7), respectively, and thus drop time-invariant variables such as contiguity (Total Borders) and region dummies. a Wald Test: different at p< 0.0000. b Wald Test: different at p< 0.0000. c Wald Test: different at p< 0.0000. d Wald Test: different at p< 0.0000. e Wald Test: different at p< 0.2260. f Wald Test: different at p< 0.4373. g Wald Test: different at p< 0.0500. h Wald Test: different at p< 0.2415. 68 Chapter 6. Case Selection Strategy: A Typical Case vs. A Deviant Case I now turn to my case studies, North Korea as a deviant case, and Taiwan as a typical case. By closely examining case studies, I hope to offer a fruitful integration of quantitative techniques laid out in Chapter 5 and qualitative techniques, a line of inquiry suggested by a number of recent studies (e.g. Bennett and George (2005); Brady, Collier and Seawright (2004); Gerring (2001, 2007); Goertz (2006); King, Keohane and Verba (1994); Ragin (2000)). While no consensus on the definition of case studies seems to have emerged, scholars largely agree that a case is “an instance of a class of events,” and a case study is a “detailed examination of an aspect of a historical episode to develop or test historical explanations that may be generalizable to other events” (Bennett and George 2005, 17). 68 Despite the “degrees of freedom” problem that case studies are inherently limited in their ability to establish causation (Campbell et al. 6), case studies can be particularly useful in explaining why a case violates theoretical predictions and to refine an existing hypothesis or specify its scope conditions (Levy 2008). Case Selection: a Typical Case Taiwan is a typical case for my theory. Chiang Kai-shek is a Power Seizer who had emerged as a revolutionary to lead China, however fragmented and vulnerable to pow- erful local warlords the country was, in the 1920s until he retreated to Taiwan after defeat by the Communists. He was the son of a petty salt merchant and his second wife, and the ascent to power was solely of his own making. Chiang Kai-shek first joined 68 Seawright and Gerring (2008, 296) define case study as “the intensive (qualitative or quantitative) analysis of a single unit or a small number of units (the cases), where the researcher’s goal is to understand a larger class of similar units (a population of cases). 69 Figure 8: Number of MIDs Divided by Years in Office by Leaders the Tung Meng-Hui, or the Revolutionary Alliance, a secret society and the forerun- ner of the Kuomintang. When revolts broke out against the Manchu forces in October 1911, Chiang Kai-shek left the Japanese Army where he had previously signed up as an officer, and took the field against the Manchu forces. He proved himself a capable commander leading a successful attack on Hangchow that led open to military posi- tions in the Shanghai area. Chiang Kai-shek ultimately came to dominate the party after Sun Yat-sen’s death in March 1925 and later was recognized as leader of his National Government at Nanking by the Western powers in November 1928. During his years as China’s leader, Chiang Kai-shek faced an uneasy country plagued by domestic strife and armed conflict with intensifying Japanese invasion that lasted throughout 1930s. With the nation in political, social and economic turmoil, Chiang Kai-shek sought political unification by force. After years of military struggle, defeat by the Communists, and the 70 consequent retreat from the mainland, Chiang Kai-shek announced that he had resumed the Presidency of China in March 1950 and held onto power in Taiwan until his death in 1975. During the two decades of his rule in Taiwan, Chiang Kai-shek had received hundreds of millions in American Military aid that helped him build a sizable armed force on the island. At the same time, Taiwan created an economic boom based on a network of textile and electronics industry using low labor costs due to the repressive policies towards trade unions. Thriving industry did not ease repressive military rule under Chiang Kai-shek. Taiwan, however, has grown into a prosperous, open, democratic society under the auspices of Chiang Ching-kuo who had been groomed to succeed Chiang Kai-shek. After decades of political repression of the native Taiwanese on the island and the con- tinuous public advocacy of a military invasion into the mainland by his father and the conservative Kuomintang, Chiang Ching-kuo initiated a series of reforms including lift- ing of martial law, liberalization of press laws, and allowing for the formation of an opposition party, the Democratic Progressive Party. These were changes consistent with what was publicly declared to be the KMT policy in the last years of Chiang Ching-kuo that the KMT policy was no longer to seek to replace the Communist regime on the mainland, but to push for “political reforms, freedom of the press, and economic liber- alization,” 69 a remarkable reversal of a long-time policy advocated by his father. Chiang Ching-kuo is known to have worked to ensure that there would not be a family dynasty or military regime that followed him 70 paving path to power for Lee Teng-hui, a native 69 Foreign Broadcast Information Service, Daily Report, China, Sept.9, 1987, p.43, citing Agence France Press report. 70 Eric Pace. (Jan. 14, 1988). Chiang Ching-Kuo Dies at 77, Ending a Dynasty on Taiwan. The New York Times. Available at <http://www.nytimes.com/1988/ 01/14/obituaries/chiang-ching-kuo-dies-at-77-ending-a-dynasty-on-taiwan.html? 71 Taiwanese and a Cornell University educated technocrat, first as the Vice President and then as his successor. Figure 8 illustrates how the number of MIDs initiated by the leader changes dramatically around the point of leadership transition from Chiang Kai-shek to Chiang Ching-Kuo. Chiang Kai-sheck initiated 20 MIDs over his 25-year term, a stark contrast to only 1 MID by his son Chiang Ching-Kuo over ten years. A typical case is also a representative case (Seawright and Gerring 2008), which can be well explained by an existing model. In this case, the puzzle of interest for the researcher lies within the case (Seawright and Gerring 2008). Choosing a case with respect to a theory involves identifying the relevant variables that have “typical” values relative to the overall causal model (Gerring 2007). Generally speaking, it may be conceivable that a case is typical relative to some causal relationship if the parameters take the mean values, although it is by no means assured. Sometimes cases with atypical scores on a particular dimension may be even more typical than cases whose values lie close to the mean (Gerring 2007, 93). For both the North Korea and Taiwan cases, I use process tracing which is “particularly useful as a supplement in large-N statistical analyses” and for “obtaining an explanation for deviant cases” (Bennett and George 2005, 214). Process tracing is an “analytic tool for drawing descriptive and causal inferences from diagnostic pieces of evidence – often understood as part of a temporal sequence of events of phenom- ena”(Collier 2011, 824) Process tracing is widely used in combination with large-N statistical studies 71 to provide additional evidence about cause and effect through em- pirical analysis of decision making “at the individual, small group, and organizational pagewanted=all> Accessed through January 2019. 71 See for example,Walter (2004); Sambanis (2004). 72 levels, including “the analysis of leaders’ perceptions, judgments, preferences, internal decision-making environment, and choices”(Levy 2008, 11). Process tracing focuses on the unfolding of events over time either relying on diverse primary and secondary sources, such as documents, memoirs, and biographies or by intensive interviewing. 72 I argue that the succeeding personalist leaders who didn’t have to fight for their supreme domestic power are less belligerent in their dealings with international affairs because their beliefs in military efficacy is bounded by their experiences while obtaining domestic power. Self selection into power characterized as belief in military efficacy and ambitious predisposition, in turn, are the key independent variables in my theory that is evaluated against the existing rival explanation of domestic institutional constraints on the executive. Thus, a crucial task to empirically establish through my case studies involve: (a) that Kim Jung-un and Chiang Ching-Kuo indeed had different paths to power compared to their respective first generation ancestor, Kim Il-Sung and Chiang Ching-Kuo in terms of competition and use of force, and the paths to power helped them develop different beliefs in the efficacy of use of force; (b) that Kim Jung-un and Chiang Ching-Kuo had different temperaments and dispositions compared to their fathers. I focus and interpret diagnostic evidence that addresses these tasks through process tracing (Collier 2011). Lastly, the selection of the deviant and typical cases appears to violate the long-standing social science wisdom that we should not select on the dependent variable (Geddes 1990; King, Keohane and Verba 1994; Brady, Collier and Seawright 2004). Selecting on the dependent variable would be indeed problematic if the researcher treats 72 For prominent scholar work using process tracing, see, for example, Tannenwald 1999 on nuclear taboo after World War II; Bennett 2010 on the Fashoda crisis, Schultz 2001 on democracy and coercive diplomacy; and Lerner 1958 on social change in a Turkish village. 73 the case as if it were “representative of a population” (Seawright and Gerring 2008; Gerring 2007). However, as (Seawright and Gerring 2008, 301) argue with extreme case method, selecting on the dependent variables is acceptable if the objective is to provide a “full range of variation as well as a more representative picture of the population” (Seawright and Gerring 2008, 301). Moreover, selecting on the dependent variable is generally acknowledged as not being a problem for process-tracing which follows an arguably different inferential logic (Bennett and George 2005; Collier, Mahoney and Seawright 2004; Seawright and Gerring 2008; Bennett and Elman 2006). Case Selection: a Deviant Case North Korea is seemingly a deviant case for my theory. My central argument that the Power Successors are more peaceful than the Power Seizers seems to be at odds with the case of North Korea, a country currently ruled by a third leader of the Kim family, Kim Jung-un. Within a decade since North Korea first launched its nuclear test on October 9, 2006, five nuclear tests have been conducted, the last of which occurred in September 2016. 73 The standard date for the start of nuclear project by North Korea is 1980, but documents released by Eastern European countries suggest that North Korea’s nuclear ambition began as early as the 1960s(Hymans 2012). Note North Korea’s leadership transition from the first generation leader, Kim Il-Sung, to his son, Kim Jong-Il in 1994, and then another transition to the current leader, Kim Jung-un in 2012. The dates of the start of the nuclear project and the subsequent tests show that the leadership transitions in North Korea have not halted, much less reversed, North Korea’s nuclear program. If anything, the nuclear project seems to have been accelerated under the leadership of 73 The first nuclear test was conducted on Oct. 9, 2006; the second on May 25, 2009; the third on Feb. 12, 2013; the fourth on Jan. 6, 2016; and the fifth on Sep. 9, 2016. 74 Kim Jung-un. 74 The growing concerns posed by North Korea’s nuclear capability challenge my theory that the belligerency by personalist leaders will become less pronounced as leadership changes through hereditary transitions. A North Korea observer opines that “North Korean strategic objectives had evolved from defensive deterrence to the ca- pability to pursue nuclear retaliation” 75 based on statements released by North Korean officials following nuclear tests in 2006, 2009, 2013, and 2016. If true, this conclu- sion indicates a growing danger of the use of nuclear weapons by the North Koreans if targeted. On the other hand, however, as alarming as the North Korean nuclear threat seems, North Korea’s pursuit of nuclear project itself might not imply much about its leader’s willingness to initiate and risk international armed conflict. At the heart of the issue, then, lies the question of whether Kim Jung-un is willing to actually use force against the United States or its neighboring countries, South Korea and Japan. Upon closer look, the records in the Militarized Interstate Disputes (MID) dataset (Palmer et al. 2015) sends a mixed message about the extent of North Korea as a deviant case for my theory. Figure 8 illustrates the number of MIDs divided by the 74 As of this writing in January 2019, there are mixed signs and interpretations about recent development between North Korea and the United States. In June 2018, a his- toric face-to-face meeting between US president Trump and North Korea’s leader Kim Jung-un took place that produced hopes for thawing of tension on the Korean Peninsula. However, North Korea is still reportedly working on missile development. As of this writing, there are reports that President Trump and North Korean leader Kim Jung-un will meet for a second time. Mark Landler and David Sanger. Trump and Kim Jong- un to Hold Second Summit Meeting Next Month. The New York Times. Available at <https://www.nytimes.com/2019/01/18/us/politics/trump-kim-summit.html> Accessed in January 2019. 75 Scott Snyder, Oct 7, 2016. http://blogs.cfr.org/asia/2016/10/07/ north-korea-ten-years-after-the-first-nuclear-test/ 75 total years in office by the leader. My theory predicts that Kim Jong-Il, the succeeding leader, initiated a fewer number of MIDs averaged over his reign than his predecessor, Kim Il-Sung, had. 76 The figure, however, indicates that over his 46-year reign, Kim Il-Sung initiated 23 MIDs, while Kim Jong-Il initiated 11 MIDs over his 17-year rule. On the other hand, a further breakdown of MIDs (Palmer et al. 2015) based on the intensity of the incidents and whether any fatalities were involved reveals an aspect of MIDs that a simple comparison was masking (Table 8): 77 Out of a total of 23 MIDs during the 46 years of leadership by Kim Il-Sung, 22 incidents were higher levels of MIDs which actually involved use of force; 11 cases involved fatalities, and only one case was of lower level of MID. On the contrary, among 11 cases of MIDs that Kim Jong-Il oversaw in a span of 17 years, six incidents involved use of force, while almost an equal number of 5 cases were lower levels of events. In other words, the rate of higher levels of MIDs, including a war, during Kim Il-Sung’s reign was much higher comprising of almost all MID incidents, while that under Kim Jong-Il amounts to only half of all MIDs. 78 The narrative so far paints an ambiguous picture about the difference in conflict propensity by Kim Il-Sung and Kim Jong-il, and warrants a closer look into the paths pursued by those two leaders when coming to power. 76 MID dataset goes only to 2010, a date before the third generation North Korea leader, Kim Jung-un, came to power in 2012. 77 Number of Total MIDs, Number of MIDs involving fatalities, Number of lower level of MIDs, Number of higher level of MIDs, Number of War initiation, Number of years in office. Total number of MIDs is the sum of lower level and higher level of MIDs. 78 For detailed narrative of the MIDs, see Appendix Table 16, 17, 18, 19, 10. 76 Table 8: Breakdown of MIDs Based on Intensity in North Korea Leader MIDs Involving fatalities lower level higher level War Tenure Kim Il-Sung 23 11 1 22 1 46 Kim Jong-Il 11 2 5 6 0 17 To say a case is deviant is to say that the case is a theoretical anomaly taking a surprising value (Seawright and Gerring 2008; Gerring 2007). Usually, the purpose of a deviant case analysis is “to probe for new – but as yet unspecified – explanations” (Seawright and Gerring 2008, 32). However, as scholars have noted it is conceivable that a close examination of a deviant case will lead to the conclusion that “the case does not violate the theory’s core predictions – because of measurement error, inappropriate operationalization of key concepts, failure to incorporate important contextual variables, recognition that the case falls outside of the theory’s scope conditions, or for other rea- sons” (Levy 2008, 13). Thus, the result may lead to save a theory from potentially harming deviant evidence. 79 As a result, the researcher either demonstrates that the case is not really deviant, or refines the existing theory. 79 Levy 2008 notes that deviant case study can contribute to hypothesis testing instead of hypothesis generating because of the unintended result of rescuing a theory from potentially damaging evidence. Gerring 2007, however, categorizes such instances as an “influential case” and distinguishes it from a deviant case. 77 Chapter 7: Taiwan: A Typical Case Figure 8 confirms my hypothesis in the case of Taiwan that the rate of conflict initiation by the Power Successor (Chiang Ching-kuo) is depressed compared to that by the first generation Power Seizer (Chiang Kai-shek). A further breakdown of MIDs (Palmer, d’Orazio, Kenwick, and Lane, 2015) based on the intensity of the incidents and whether there has been any fatalities involved (Table 9) also supports the claim. Out of 20 MIDs during Chiang Kai-shek’s 25-year reign, 3 incidents involved fatalities of more than one and 18 events were higher levels of MIDs such that accompanied actual use of force. Only two events were lower levels of MIDs that are coded as either simple “display of force”, “threat to use force”, or “no militarized action”. On the contrary, there is only one case of the higher level of MID marked under the 10-year rule of Chiang Ching-kuo. Table 9: Breakdown of MIDs Based on Intensity in Taiwan Leader MIDs Involving fatalities lower level higher level War Tenure CKS 20 3 2 18 0 25 CCK 1 0 0 1 0 10 The purpose of examining a typical case such as Taiwan is to identify the mechanism under which the identified independent variables (Power Seizer vs. Power Successor) lead to the predicted outcome in the unfolding of events. To this end, I focus on the two mechanisms outlined in my theory chapter (Chapter 3). First, did Chiang Kai-shek and Chiang Ching-kuo indeed have different paths to power and were exposed to different types of events that helped them develop varying levels of beliefs in the efficacy of use of force? Second, did Chiang Kai-shek and Chiang Ching-kuo have different temperaments and dispositions that contributed to their inclination to conflict 78 initiation? Mechanism 1: Divergent Paths to Power The first mechanism sheds light on the contrasting courses of power attainment po- tentially experienced by Chiang Kai-shek and Chiang Ching-kuo. The violent path to power beset by bloody domestic strife against powerful warlords and the Communists, as well as fights with the Japanese invaders have likely contributed for Chiang Kai-shek to develop practical skills necessary to command military forces in the field. Over time, the practice of military force not only formed his perception that force is useful and effective, but hardened into a routine and legitimate instrument of politics when con- fronted with challenges. The account of Chiang Kai-shek’s claim to power intends to focus on his greater beliefs about his capacity to command force and the usefulness of force, and how these beliefs contributed to his apparent willingness to resort to military force, which is more pronounced than leaders whose path to power was relatively less forceful. On the contrary, Chiang Ching-kuo inherited from his father the absolute power and the domestic political conditions that feed into it, and his experiences with military command had been largely limited during his political rise. While his years as an heir had been widely plagued by internal struggles against the powerful local factions, the Communists, and fights against the Japanese, unlike his father, Chiang Ching-kuo was never an active field officer in charge of carrying out military operations against enemy forces. Not only did Chiang Ching-kuo lack the firsthand experience with force but it is less likely that he had developed a belief about the usefulness and legitimacy about military force as a political tool. 79 Chiang Kai-shek: An Undiscouraged Old Soldier Chiang Kai-shek was born in October 1887, at Fenghua, Chekiang Province, 100 miles south of Shanghai, in the midst of a near-anarchy following the breakup of the Manchu Dynasty. He was the son of a petty salt merchant and his concubine, and had a grim childhood which he recalled on his 50th birthday: "My father died when I was 9 years old. The miserable condition of my family at that time is beyond description. My family, solitary and without influence, became at once the target of much insult and abuse. . . . It was entirely due to my mother [a devout Buddhist] and her kindness and perseverance that the family was saved from utter ruin. For a period of 17 years - from the age of 9 until I was 25 years old - my mother never spent a day free from domestic difficulties." 80 Though from a humble background, Chiang Kai-shek managed to enter the highly competitive National Army School near Peking in 1907, known as the Paoting military Academy. As one of the few Chinese cadets, he was later selected to go to Shimbu Gakko, a Japanese Military Academy. His desire to follow a military career in Japan was known to have been inspired by a series of events where the Japanese demonstrated an industrial base and a navy capable of defeating Russians out of Chinese cities – Port Arthur and Mukden, an impressive feat in the eyes of many Chinese who had been accustomed to view Japan as a mere traditional society for a long time. It was then in Japan that Chiang Kai-shek became involved with the Revolutionary Alliance, a secret society and a forerunner of the Kuomintang, against the Qing Dynasty led by Sun Yat-sen (Boorman, Cheng and Krompart 1967, 200). Chiang Kai-shek’s ties and engagement with the Revolutionary Alliance al- 80 Alden Whitman. (April 6, 1975). The Life of Chiang Kai-shek: A Leader Who Was Thrust Aside by Revolution. The New York Times. Available at <http://www.nytimes. com/learning/general/onthisday/bday/1031.html> Accessed through January 2019. 80 lowed him to become exposed to numerous military combats from the age of twenty- four. When the Revolutionary Alliance launched an uprising in the city of Wuhan in 1911, Chiang left his army base in Japan and joined the movement in Hangchow where he boldly and successfully led an assault on the headquarters of the Governor Gen- eral. After the seizure of Hangchow, Chiang turned down the position of governor of Chekiang and joined as a ranking military officer in the revolutionary ranks based in Shanghai (Tong 1953, 33). Over the next ten years Chiang strengthened his reputation as a courageous, dedicated and self-disciplined soldier. With the aid of the Soviet in 1926, Changsha and the three Wuhan cities finally fell to the Revolutionary Army led by Chiang. He was constantly engaged in military operations of one sort or another including conventional assaults, guerrilla attacks, and terrorist strikes, and usually had to live undercover with a price on his head. Sun Yat-sen also clearly thought Chiang Kai-shek was an invaluable soldier even as Chiang developed a reputation as a stubborn and often erratic officer, a personality trait that often got him into trouble with other officers (Taylor 2009a, 30). During the course of the revolution, Chiang Kai-shek had been a soldier, a clandestine agent, and even an assassin a few times. Although uncon- firmed he was reportedly also involved in at least one robbery conducted on behalf of the revolution (Taylor 2009a, 12). As Chiang Kai-shek’s reputation and status strengthened within the Revolu- tionary Alliance and later within the Kuomintang (KMT), so did the political challenges against his leadership. In 1925 Chiang Kai-shek became a garrison commander at Can- ton. As dedicated a military officer as Chiang seemed, he was still not considered as a potential contender for the top leadership within the KMT by many observers by then. In August 1925, however, with the assassination of Liao Chung-k’ai, Chiang Kai-shek suddenly became the only senior military leader of the Revolutionary Army, the mili- tary arm of the KMT, and one of the most powerful political figures within the KMT. 81 Finally, in the summer of 1926, Chiang Kai-shek became Commander-in-Chief of the National Revolutionary Army, and Commander-in-Chief of the KMT in January 1928. Chiang’s rise to power, however, did not come without political backlash, and even after he was appointed Commander-in-Chief of the KMT, his leadership position was still far from secure. Not only did his forces suffer a defeat at Hsuchow at the hands of hostile warlords in 1927, but internally, he was consistently vulnerable to political maneuvering launched by the left KMT members and other KMT factions such as one led by Wang Ching-wei. As a tactic, Chiang Kai-shek often sought to escape political gridlock by “re- signing.” For example, in August 1927, when the infighting between the KMT factions in Nanking and Wuhan intensified challenging Chiang’s leadership, he made another resignation. In reality, Chiang knew that without him the Nanking regime could neither raise funds to maintain itself nor launch the Northern Expedition. While his repeated resignations were largely based on his sense that he was indispensable, those moves had often given him a much-needed political break as well as an unusual reputation for rectitude. In 1928, for the first time since 1911, the Kuomintang became the recognized authority throughout 95 percent of China from the Ussuri River in the north to the border with Southeast Asia. Over the years Chiang Kai-shek consolidated his power as the leader of modern China and as one of the most important statesmen on the world stage. By 1930, the Nanking-based Republic of China (ROC) had been recognized by most of the world’s governments as the legal authority of all China, and Chiang Kai-shek as the leader of all China. Chiang Kai-shek initially chose to seek political unification by force of arms and consolidation of KMT power within China in precedence to rebuffing the ever growing Japanese threat in China. The Communist bands were largely scattered in the mountain areas and in Kiangsi and neighboring provinces. While the KMT power was 82 dominant compared to the small Communist faction (CCP), as early as 1927 since the “cleansing” against the Communists led by the KMT’s Supervisory Committee, Chiang considered the removal of CCP as a priority in light of their international links with the Soviets. In turn, Chiang’s failure to resist the Japanese provoked enormous criticism against him which led to yet another resignation in 1931 – only to head back to Nanking after letters and telegrams poured in beseeching his return – as well as the Hsian crisis in 1937 in which Chiang Kai-shek was held captive by Chang Hsueh-liang, a strong war- lord in Manchuria often referred to as his nickname, “Young Marshall”, who demanded that Chiang form a popular front to resist the Japanese and “save the country” (Crozier and Chou 1976, 182). The Hsian crisis, however, only strengthened Chiang Kai-shek’s stance within the KMT and the alliance with CCP as the anxious CCP leaders, Mao and Zhou, as well as Stalin who feared the prospect of a pro-Japanese KMT regime in Nanking without Chiang called for his release (Taylor 2009a). Despite initial reluctance, Chiang Kai-shek got engaged in the all-out war with Japan in July 1937 at the Marco Polo Bridge outside Peking. The war efforts against the Japanese were then replaced by civil strife against the Communists that erupted afresh as the 1946 cease-fire accord was quickly breached after the defeat of Japan. Chiang Kai-shek at first seemed to have the upper hand with his three million troops to Mao’s one million, and more than $3 billion in American aid poured into his regime between August 1945 and the middle of 1948. 81 Moreover, Chiang believed that he could prevail against the Communists by inducing Soviet cooperation for which he placed his son, Chiang Ching-kuo to be in charge of. However, by the spring of 1947 it became clear that the initiative favored the Communists. In December 1949, Chiang fled to Taiwan 81 Alden Whitman. (April 6, 1975). The Life of Chiang Kai-shek: A Leader Who Was Thrust Aside by Revolution. The New York Times. Available at <http://www.nytimes. com/learning/general/onthisday/bday/1031.html> Accessed through January 2019. 83 declaring Taipei to be the temporary capital of China, and resumed the Presidency of China in March 1950. Just when it seemed like Chiang Kai-shek’s fortune was fading away, he was rescued by the outbreak of the Korean War in June 1950 and the participation of the Soviet Union on the side of the Communist North Korea. The United States quickly bolstered the Taiwan as a counterweight to the Communist forces in Asia pouring polit- ical, economic and military assistance into the island. America sent its Seventh Fleet to the Strait of Taiwan, and began to train about 600,000 men in the Nationalist army. In 1954, the United States and Chiang Kai-shek signed a mutual defense treaty. Embold- ened by a series of international events that at least initially seemed to secure his position both internationally and domestically, Chiang Kai-shek never gave up the thought of re- turning in triumph to the mainland. The account of Chiang Kai-shek’s life thus far suggests his firm beliefs about the usefulness and legitimacy of force as a political tool as he learned from his first hand experiences as a soldier and military commander first in the participation of the Revolu- tionary Alliance against the Dynasty and then in the struggles against the Japanese and the Communist from his youth onward. Coming from a humble background, his dra- matic rise to political leadership within KMT had been constantly challenged and threat- ened by different factions, but Chiang managed to maintain his prevalence through force and action, including mass extermination of the Communists in 1927. Chiang’s belief about use of force continued to persist as he insisted that the only way to bring peace in Asia is the destruction of the Communist regime in Peking even after the dismal retreat from the Mainland in the late 1940s. In 1966 at the inauguration of a fourth term presi- dency, Chiang called himself an “undiscouraged old soldier,” and vowed that he “would exterminate Mao Tse-tung and his cohorts, liberate our mainland compatriots and es- 84 tablish on the ruins a new country of unity and freedom.” 82 Although never realized, Chiang Kai-shek called for a nuclear Taiwan approving establishment of a clandestine nuclear weapons research laboratory in Taiwan, as well as instructed secret planning for a surprise attack on the mainland. Around the time when Chiang Kai-shek was sworn in for his fifth six-year term as president in 1972 and when he shortly afterward fell ill, his international stature began to erode in front of his eyes with the Communists in Peking being voted into the world organization and President Nixon establishing an entente with Peking followed by other countries recognizing Peking as the legitimate Chinese authority, Japan among others. Chiang, however, never admitted defeat until the last day of his life. Chiang Ching-kuo: A Competent Administrator and Scholar Brings Reform to Taiwan Chiang Ching-kuo, the son of Chiang Kai-shek with his first wife, and the only biological son, had a very different path to ascendancy in every aspect throughout his political career compared to his father. Though initially neglected in his first few years of childhood by his father, Chiang Ching-kuo had been well groomed as the next line of successor from early on his youth. Chiang Kai-shek cared for Ching-kuo’s ed- ucation, personally arranged tutoring and permitted his son to go study in the Soviet Union at the age of 15. Chiang Ching-kuo was a fine scholar and student in military tactics, but unlike his father, the younger Chiang was never a soldier or an active field officer and had virtually no combat experiences. While his father had to constantly 82 Alden Whitman. (April 6, 1975). The Life of Chiang Kai-shek: A Leader Who Was Thrust Aside by Revolution. The New York Times. Available at <http://www.nytimes. com/learning/general/onthisday/bday/1031.html> Accessed through January 2019. 85 stave off political maneuvering and even physical threats during his rise to power, Chi- ang Ching-kuo’s position had been almost ironclad under the stewardship of his father. While Chiang Kai-shek stubbornly insisted on the destruction and replacement of the Communist regime in Peking even after the retreat to Taiwan, Chiang Ching-kuo, who was more objective in assessing world trends and Taiwan’s stature, called for a more fine-tuned plan centered on the expectation that the successful economic and cultural model in Taiwan would eventually influence political developments on the mainland. Chiang Ching-kuo was born in the village of Chikow to Mao Fu-mei, Chi- ang Kai-shek’s first wife. In the early years into his childhood, Chiang Kai-shek didn’t show much interest in the boy in part because of his estranged relationship with Fu- mei whose illiteracy and traditional demeanor often embarrassed him. However, around Ching-kuo’s tenth birthday, Chiang Kai-shek began to show interest in him suspecting that Ching-kuo would be his only biological heir. Chiang Kai-shek personally arranged teachers and tutoring for his son, and emphasized classic Chinese literature as well as Chinese calligraphy. Ching-kuo’s teacher reported to his father that the boy was “not brilliant, but like to study very much”(Mao 1965), and Chiang Kai-shek believed that his son was a “teachable” child(Ching-kuo 1989). In 1925 when Ching-kuo turned fif- teen, Chiang Kai-shek first sent him to a small academy combining classical and modern approaches to education that his old friend, Wu Chih-hui opened in Peking, and soon afterwards to the Soviet Union at Sunovka (Sun Yat-sen University) which was during a short-lived period of cordial relations with the Soviets. There Ching-kuo was taught Russian language and introductory courses on communism and imperialism, as well as practical skills such as government infiltration and how to create peasant and labor movements. Ching-kuo went on to join a Communist Youth organization and became a Red Army cadet for a time. In his first years Ching-kuo studied military tactics, admin- istration, transportation, topography, principles of artillery, and guerilla warfare, as well 86 as numerous courses in Marxism-Leninism and the techniques of political commissar work in the armed forces (Cline 1989, 16). After his father turned against the Chinese Communists in 1927 purging a large number of them in Shanghai, Chiang Ching-kuo in turn was held hostage for po- litical reasons in the Soviet Union for nearly ten years afterwards until his return to China in 1937. During these years, Ching-kuo’s jobs included stints as a village worker in Zhukova on the outskirts of Moscow and toiling in a heavy industry installation in Uralmash. Though a captive, as the son of the Chinese leader, there is little sign that Chiang Ching-kuo was severely mistreated by the Soviets. However, the years in the Soviet Union had had a lasting impact on Chiang Ching-kuo. Chiang Ching-kuo later accounted that witnessing the Stalinist purges in those days made him a staunch enemy of Communism. 83 Later in 1941 when the Soviets overtly courted the Japanese, Ching- kuo became more outspokenly anti-Communist. On the other hand, Chiang Ching-kuo is also believed to have a strong Marxist slant. In what is a rare case among KMT officials, he often mingled with the people to listen to them and to understand their hardships, and his reforms carried out in his early days in Taiwan had semblance to those overseen by his old Moscow schoolmate, Deng Xiaoping on the mainland. 84 While Chiang Kai-shek’s rise to power was ultimately enabled by his courage and skills in action as a soldier and a commander demonstrated in numerous battlefields 83 Eric Pace. (Jan. 14, 1988). Chiang Ching-Kuo Dies at 77, Ending a Dynasty on Taiwan. The New York Times. Available at <http://www.nytimes.com/1988/ 01/14/obituaries/chiang-ching-kuo-dies-at-77-ending-a-dynasty-on-taiwan.html? pagewanted=all> 84 Yang Hengjun. (December 10, 2014). Chiang Ching-kuo, China’s Demo- cratic Pioneer. The Diplomat. Available at <http://thediplomat.com/2014/12/ chiang-ching-kuo-chinas-democratic-pioneer/> Accessed through January 2019. 87 especially during his early career, Chiang Ching-kuo, the only son of the Commander- in-Chief, never had the same need to prove his capacity as military officer as his father did to either earn or secure his status within the KMT. In fact, though drilled in warfare as an academic subject, Ching-kuo never was actively engaged in any sort of a battle as a combatant himself. Instead, over the years since the young Chiang was released from the Soviet Union in 1937, he was groomed by his father in various non-combative posts, mostly in education, youth work, and internal security. To start with, upon his return to China in 1937, Chiang Ching-kuo was appointed deputy director of the Provincial Peace Preservation Corps, and then became a member of the new Youth Corps central commit- tee and director of its Kiangsi branch. From then on, Ching-kuo had been extensively involved in military education and youth training. Under his leadership, a large Youth Cadre School was established in Chingking in 1943, and Ching-kuo oversaw thousands of youth leaders being trained in the new school whose curriculum was designed by Ching-kuo himself from scratch. The graduates from the Youth Cadre School were edu- cated, patriotic, and idealistic who came to regard Chiang Ching-kuo as a devoted leader they could believe in. In addition, Chiang Ching-kuo initiated to create Educated Youth Expeditionary Army in 1944 with the purpose of creating a tough and truly effective corps of commissars. The young leaders later became staunch followers and political allies of Chiang Ching-kuo for years to come. Chiang Ching-kuo also took up the responsibility to build an economic model in a poor area known as Kannan in 1939 where he worked to improve the condi- tions along the “five haves”: sufficient food, clothing, shelter, books, and work (Taylor 2009b). The work in Kannan gave Ching-kuo a reputation as a competent and incor- ruptible leader with experience in administration and local development. Though unsuccessful, Chiang Ching-kuo was initially involved in diplomacy and foreign policy with the Soviets at the end of the World War II when Manchukuo was 88 cleared of Japan’s Kwangtung Army and its allies. Chiang Kai-shek appointed Ching- kuo Special Foreign Ministry for Manchuria in hopes of accommodating and winning over the cooperation of the Soviets in its competition against the Communists. While Ching-kuo’s failed policy towards the Soviet Union did stir strong criticism among Chi- ang Kai-shek’s enemies within the Party on his decision to assign his son the responsi- bility in this area, both the elder and the young Chiang remained unscathed politically. After moving to Taiwan in 1949, Ching-kuo started to get extensively involved in internal security and secret police operations. Soon a political Action Committee was established in Kaohsiung in an effort to coordinate numerous intelligence operations that came in droves to the island. In 1950, Chiang Kai-shek formally appointed his son the director of General Political Department of the Ministry of Defense with the responsibility of restructuring the political officer corps including military surveillance system, counterintelligence, and reinstallation of KMT activities throughout the military (Ching-kuo 1989, 154). The restructuring gave Chiang Ching-kuo the unrivaled power over civil as well as military officers (P’eng-fei 1994), and he had all the various wings of the intelligence and secret police networks pulled under his control. By 1954, he acted as chief of the National Security Bureau (NSB), the central intelligence coordinating body, and was the deputy chief of the 6th Division of KMT headquarters, another body that dealt with security and intelligence matters. Eventually, Chiang Ching-kuo became Minister of Defense in 1965 upon appointment by Premier Yen. While Ching-kuo’s qualification itself was never a subject of dispute, his smooth, undisturbed, bloodless political ascendancy owes entirely to his father. Although Chi- ang Kai-shek had at least two formal marriages and extramarital relations with numerous mistresses, Ching-kuo was his only biological son. Chiang Kai-shek’s only other off- spring, Chiang Wei-kuo, was an adopted son who was not interested in politics. As the only son of the mighty Generalissimo, Ching-kuo faced little competition. Without 89 his father, the young Chiang, however competent, would not have the opportunity to put his own stamps on various key government posts in areas as extensive as education, military, foreign policy, and intelligence from the young age. The young Chiang was groomed to succeed his powerful father, and ultimately became deputy premier in 1969 to assume a formal leadership role in the overall administration of the government with- out any resistance within the party. When Chiang Kai-shek decided it was time for his son to formally take charge in 1972, then Premier Yen conveniently stepped down from the premiership making way for the younger heir’s succession. After the death of Chi- ang Kai-shek, C.K. Yen again conveniently decided to step down from the presidency in 1978 recommending Chiang Ching-kuo as the successor. The divergent paths to power had led the two leaders to greatly differ in their approaches to the Communist Party on the mainland as well as towards the domestic population on the island. Unlike the elder Chiang who had remained authoritarian and repressive throughout his entire life, Ching-kuo understood that it was necessary to gar- ner support from the native Taiwanese population for the Government. After becoming President in 1978, Ching-kuo put into effect a series of liberalization measures including ending the martial law and allowing for the formation of an opposition party, the Demo- cratic Progressive Party (DPP). In 1986 at the behest of Chiang Ching-kuo, the Standing Committee established a new political reform committee whose task was to expedite po- litical reform and publish a timetable for the public to understand the reformist direction the Party was taking (Wang 1995). In the first free, democratic parliamentary elections in Taiwan in December 1986, the three-month-old DPP had won 23 of the 44 seats it contested (Taylor 2009b). The younger Chiang’s approach towards the mainland differed from that by his father as well. Chiang Ching-kuo believed that Taiwan should continue making progress towards a full-fledged democracy and become an example for China to induce 90 changes on the mainland rather than follow the largely military and confrontational path advocated by the older Chiang. By 1980, Taiwan approved people-to-people contact with the mainland for the first time. Ching-kuo himself discounted his father’s increas- ingly ritualistic calls for a military attack on the mainland, and expected that Taiwan’s economic and cultural strengths would eventually influence political events in China. Mechanism 2: Like Father, Like Son? Personality Traits and Foreign Policy The second mechanism focuses on the disposition and personality traits in- herent in the Power Seizers and Power Successors. My argument builds on the no- tion that the Power Seizers who came to power through forceful means tend to possess greater international ambitions and risk tolerance. The same traits that helped leaders to seize power domestically also lead them to alter the status quo in the international arena through force. As a result, the Power Seizers tend to be more conflict prone than typical leaders. Chiang Kai-shek, for example, who first came to power as a revolu- tionary against the Qing Dynasty, didn’t hesitate to take the field against enemies or assume even more dramatic missions such as assassination. During his exile in Taiwan, he never ceased to seek replacement of the Communist regime on the mainland, and had remained vocal about his hostile policies towards his enemies. Moreover, previous research shows that personalist dictators have insatiable need for attention and dominance over others which fuels their desire for international conquest and leads them to take upon risky path (Glad 2002). As will be detailed in the following section, several personal accounts about the elder Chiang as well as his personal diaries reveal that Chiang Kai-shek viewed himself as extravagant and arrogant. On the other hand, the Power Successors may lack the international ambitions and revisionist world view that served as the driving force for their fathers to emerge 91 in their domestic struggles because they didn’t have to seek the power in their hands in the first place. As discussed earlier, Chiang Ching-kuo, never had to use force or prove his ability as a capable combatant during his entire political career. He was certainly a competent administrator, but he was never a combatant. Thus, when his father was bitter and furious about a series of diplomatic recognitions of China over Taiwan by major powers starting in the 1960s, Chiang Ching-kuo rather called for a non-militaristic plan for the CCP understanding the compelling interests of the major powers in establishing good relations with China. Chiang Kai-shek: Provocative, Stubborn, and Narcissistic, Climbs to the Top Lead- ership Chiang Kai-shek’s rise to power during the revolution owes much to his un- matched self-assurance, willpower, and decisiveness. His superiors saw in Chiang a steel will of commanding personality, and Sun Yat-sen, in particular, clearly found Chi- ang’s courage, incorruptibility, and decisiveness a valuable asset. The same traits, how- ever, also continued to cause Chiang trouble with his colleagues and subordinates. In 1920 Chiang was constantly on the move with his battalions and Sun Yat-sen winning one battle after another against the southern warlords. When the pressure of combat, however, got to him and he frequently lashed out at his officers, Sun Yat-sen remarked in a letter to Chiang that his “fiery temper” and “hatred of mediocrity” is too exces- sive and often created quarrels and made cooperation difficult (Loh 1971, 52). Chiang himself was well aware of his quick temper and he continued to chastise himself for displaying anger not only to his fellow officers and subordinates but also towards ordi- nary people like rickshaw pullers (Loh 1971, 33). In addition, Chiang had a myriad of traits he thought were flaws in his personality, and often detailed them in the diary that he began keeping since 1918. He complained in his diary about his being “"ruthless and tyrannical; irritable; conceited; stubborn; wicked;. . . extravagant; jealous; stingy; 92 lascivious; arrogant; full of sorrow and indignation” were problems (Taylor 2009a, 30). This frequent reflection of his own shortcomings seems in part an expression of rote neo- Confucian self-cultivation and ever-higher standards of behavior(Taylor 2009a, 90). Chiang Kai-shek’s provocative and stubborn temperament was pronounced in his encounter with General Joseph W. Stilwell who was President Roosevelt’s choice to be commander in chief of Chinese and American forces in China and sent to Chungk- ing in 1942. Stilwell was a Chinese speaker and West Point graduate who had served in China in the 1920s and 1930s. He was then widely considered as the most com- petent China specialist in the U.S. military. Like Chiang Kai-shek, he was a man of action with strong willpower, but his relationship with Chiang soon proved to be less than cordial. The outspoken General and the stubborn Chiang soon fell out. In his re- port to the War Department, Stilwell remarked that Chiang sought to “dominate rather than unify and lead” China against Japan. 85 In 1944, President Roosevelt proposed that Stilwell be given command of the Nationalist troops for the effective overthrow of the Japanese military. However, Chiang interposed conditions such as Lend-Lease control and effective authority over the General in the matter of strategy and tactics. Chiang’s at- titude infuriated Stilwell and he described Chiang as “a stubborn, prejudiced, conceited, despot” and repeatedly called Chiang “the little dummy” or “Peanut” in front of his own American staff (Stilwell 1991, 137). For Chiang’s part, he expressed grievances against Stilwell for “recklessness. . . insubordination . . . contempt and arrogance,” saying that the U.S. had rather sent him a “king of the China Theater” than a chief of staff.(Kai- shek 1942) In the end, Chiang formally demanded that General Stilwell be recalled, and President Roosevelt acceded. Stubbornness in fact was widely assumed to be one 85 Alden Whitman. (April 6, 1975). The Life of Chiang Kai-shek: A Leader Who Was Thrust Aside by Revolution. The New York Times. Available at <http://www.nytimes. com/learning/general/onthisday/bday/1031.html> Accessed Through January 2019. 93 of Chiang Kai-shek’s principal attributes which he deliberately projected to the world to induce concession. At the same time, although often introverted, he could also be a self-righteous and arrogant person. He was susceptible to flattery and praise and also gestures of respect and goodwill. In the 1920s, one Russian adviser said of Chiang “con- ceited, reserved, and ambitious” but concluded that much could be obtained from him if he were “praised in a delicate manner” and dealt with “on the basis of equality and never showing that one wants to usurp even a particle of his power” (Taylor 2009a, 5). His self-righteous and somewhat melancholic personality led him to see himself as a martyr who was destined to suffer whenever in the face of a serious setback. For example, in 1949 ahead of defeat by the Communists, Chiang believed that if he could survive in his retreat, the world will eventually turn around to his advantage. In sum, Chiang Kai-shek’s provocative temperament, stubbornness and self- righteousness served him both as an asset in confronting his enemies and hardships while caused him frequent quarrels and troubles. His ambitions and strong-willed char- acter contributed much to his tenacious resistance as a revolutionary, and then in his tireless struggle against the Communists even at times when it became clear that the enemy had turned insurmountable. Chiang Ching-kuo: A Unpretentious, Practical Leader Changes Course of Taiwan Evaluations of Chiang Ching-kuo’s characteristics are controversial with the enforcement of White Terror in Taiwan that arrested tens of thousands of people and killed at least 1,200 people between 1949 and 1992. 86 Before 1949, Chiang Ching-kuo was viewed as incorruptible, energetic, determined, amiable, pragmatic, unassuming, 86 Cindy Sui. (March 13, 2016). Taiwan Kuomintang: Revisiting the White Terror Years. BBC News. Available at <http://www.bbc.com/news/world-asia-35723603> Ac- cessed Through January 2019. 94 and someone who actually cared for the common people. While his enemies tended to call him idealistic, naïve, and left-leaning in public, they acknowledged in private that he was a good man (Taylor 2009b, 210). However, as he became identified behind the White Terror and a series of suppressions in the 1950s, people who otherwise praised his honesty and concern for the ordinary people, held Ching-kuo personally responsible for the terror. Chiang Ching-kuo and his associates themselves justified the White Terror as a necessary tool to preserve the Chinese unity in the face of serious threats of subversion. If left unchecked, they argued, the latent resistance would have taken over and called for even far worse suppression. On the other hand, the “Reynolds Case” in 1957 shows how Chiang Ching-kuo cared about his people. When Master Sergeant Robert Reynolds was found innocent at an American military court after he shot and killed a Chinese man who he thought peeped at his wife during shower, there was an immediate uproar among the Chinese people. Permitting demonstration of the crowd in front of the American embassy, Ching-kuo firmly ordered never to shot down any citizens on the streets (Taylor 2009b, 238). While he had overseen numerous arrests, detentions and even executions in his early years since he took responsibility for security on the island, his Nationalist military never fired on a public demonstration even once. Overall, Chiang Ching-kuo was known as amiable and unassuming to many foreigners as well as to Chinese. Unlike his father who was associated with his stub- bornness and quick temper, Ching-kuo was an obedient and happy child who grew up in a tranquil home full of women including his mother, grandmother, maternal great grandmother, and the servant. However, there were a few common traits father and son shared. Both the younger and older Chiang are known for their frugality, hard work, and incorruptibility. Chiang Kai-shek was famous for retaining a modest and ascetic lifestyle while leading a highly comfortable and secure life surrounded by servants and guards. However, while himself a modest man, Chiang Kai-shek tended to tolerate cor- 95 ruption and ineptness of his generals for the sake of loyalty unity of his armed forces. Chiang Kai-shek believed that corruption is best addressed once peace and unity has been achieved (Taylor 2009b, 222). Like his father, Ching-kuo proved to be an incorruptible leader whose inter- nal and spiritual drive led him to work tirelessly from early morning until mid night. However, he was someone who would try to set a model for incorruptibility and mod- esty through action. During his time in Kiangsi province after his return to China in the 1930s, Chiang Ching-kuo came to deeply understand how corruption and factionalism within the KMT started to afflict officers, military conscripts and ordinary people. As even his critics acknowledge, Ching-kuo actually cared for the common people and his heart was with these men and women. In 1924 when he was just four- teen, Ching-kuo sent a letter to his father proposing that the Wushan grammar school that he attended set up a free night school for the common people which the older Chiang rejected. Ching-kuo tried to maintain contact with ordinary people even after his power and status increased by often adventuring out to the countryside by himself. From 1950 to 1978 until he became President, he made unannounced visits to 320 towns and vil- lages, eating simple meals with the locals listening to their stories and talking to them (Taylor 2009b, 226). According to a correspondent with the US News and World Re- port, Robert Martin, Chiang Ching-kuo dared to travel around without security and was “totally unpretentious, frugal, modest, and informal,” and displayed none of “the ob- sequiousness which usually is found among top Nationalist officials” (Taylor 2009b, 224). Indeed, many of his critics who meet Chiang Ching-kuo found their conceptions of his character greatly change in the face of his affability and lack of presumption and self-importance. Moreover, Chiang Ching-kuo had more perceptive and objective eyes than his father when it came to global trends and Taiwan’s stature in the world. While Chiang 96 Kai-shek remained in his bitterness when the world started to turn around and estab- lished relations with the Communist regime in China, Ching-kuo understood that Amer- ican and Communist China simultaneously found a compelling interest in creating good relations with each other. Learning to discount his father’s call for a military attack and replacement of the Chinese regime on the mainland as a ritualistic call, Ching-kuo in- stead specifically rejected an invasion understanding that the international situation was unfavorable and that an attack would only strengthen the Communists. In sum, being an unassuming and practical leader, Chiang Ching-kuo differed greatly in his approach to the changes of world situation from his father, and understood his position and the best way to adjust to the current of military and diplomatic events. 97 Chapter 8: North Korea: A Deviant Case As examined earlier, North Korea seems to contradict my hypothesis in that the rate of conflict initiation by the Power Successor, Kim Jong-Il, is higher when di- vided by the years in office compared to that by his father, Kim Il-Sung, the Power Seizer (Figure 8). At the same time, a further breakdown of MIDs (Palmer et al. 2015) based on the intensity of the incidents and whether any fatalities were involved reveals that the rate of higher levels of MIDs, including a war, was much higher during Kim Il-Sung’s reign than that under Kim Jong-Il (Table 8). These two descriptive statistics paint somewhat ambiguous picture of the conflict propensity by these two leaders. Recall that North Korea is a “nonconstitutional” hereditary regime for my the- ory in which I hypothesize that the succeeding leaders face more political challenges and hardships from his political opponents and thus tend to be more risk-tolerant than their counterparts in the constitutional monarchies. “Nonconstitutional” cases create a scope condition for my theory where the effects of the identified independent variables (Power Seizers vs. Power Successors) might be mitigated by certain contextual variable (nonconstitutional hereditary regimes). Nevertheless, my theory predicts that the Power Successors in such nonconstitutional hereditary regimes, too, should be less conflict prone compared to their first generation Power Seizers. The aim of a deviant case is to explain “why observed outcomes do not fit the theory” (Levy 2008, 8). Thus, like the typical case of Taiwan, I focus on the two mechanisms that drive the identified independent variables (Power Seizers vs. Power Successors) to cause the predicted outcome (MID initiation). This will allow me to examine why there have been no clearly fewer conflict initiations under Kim Jong-il than by Kim Il-sung as my theory predicts. Since 2011, North Korea is run by a third generation leader, Kim Jung-un, 98 who is preceded by his father, Kim Jong-il. Because my large-N analysis introduced in my dissertation excludes the MID initiation under Kim Jung-un, and includes those under Kim Il-sung and Kim Jong-il only, I intend to focus on the trajectory paths by Kim Il-sung and Kim Jong-il. A careful and detailed examination of a deviant case will lead a researcher to either modify and refine the existing theory or to conclude that the case does not violate the theory’s core predictions after all (Levy 2008, 13). I again explore the two questions that concern how the Power Seizers and Power Successors differ: First, did Kim Il-sung and Kim Jong-il follow different types of paths to power and experience different events that helped them shape varying levels of beliefs about use of force? Second, how different are the temperaments and dispositions of Kim Il-sung and Kim Jong-il that helped them to rise to power and to initiate a conflict abroad? Mechanism 1: Self Selection into Power? The first mechanism sheds light on the potentially contrasting courses of power attainment by Kim Il-sung and Kim Jong-il. Recall that my theory predicts that the Power Successors will be less conflict prone than the Power Seizers because they inher- ited the absolute power from their predecessor rather than fought for it. The fact that there had been no fewer conflict initiations by Kim Jong-il than by Kim Il-sung leads us to closely examine the case and conclude one of the following: either the theory needs to be modified or the theory still holds but needs further scope conditions(Levy 2008). As discussed in Chapter 3, North Korea constitutes a “nonconstitutional” case where hereditary succession lacks traditional and constitutional ground to take place. When Kim Jong-il rose to power, there is no doubt that he wielded absolute power as the son of the country’s supreme leader, yet, there was no knowing that he would become the next leader. During Kim Il-sung’s reign, there had been no precedent set for hereditary succession in North Korea, and even when surmising that one of Kim Il-sung’s clos- 99 est family members was to take on the next leadership, competitions abounded among Kim Jong-il, his uncle, Kim Yong Ju, and his step brother, Kim Pyong-il and Pyong- il’s mother, Kim Song-ae. Amidst his own uncertain future, there are signs that Kim Jong-il struggled for his own succession and placed enormous effort in the selection and succession process(Jung 2014; Lee 2005). A closer examination of Kim Jong-il’s suc- cession case reveals that he himself was very much an ambitious man who ultimately struggled to solidify and retain the supreme power his father had held. In other words, in a case such as North Korea’s Kim Jong-il, where the Power Successor also “selects himself into power,” the effect of my independent variable might be mitigated. In sum, unlike the successor in other hereditary regime, Kim Jong-il did struggle to inherit the absolute power from his father compelling him to constantly prove himself as a strong and effective leader. Following the Korean War, Kim Il-sung purged various political factions from power that were part of the liberation forces during the Japanese reign and contributed to forming the North Korean regime. The members in the South Korean Worker’s Party were first to be removed, followed by the Yenan faction, the Soviet Faction, and then the opponents within his own partisan. Kim Il-sung secured absolute power by eliminating opponents who once fought alongside with him as guerrillas. After his political oppo- nents were eliminated, his dictatorship based on his personality cult started to intensify. The political purges of those factions by Kim Il-sung to consolidate his ab- solute power were not separate from securing the younger Kim’s position in the party. Kim Jong-il actively led the purges of his father’s opponents within the party and the military in the 1960s both to show his loyalty to his father and also to weaken his un- cle, Kim Yong-ju’s base who was deemed to be a strong candidate for the next North Korean leader at the time. By rooting out Kim Yong-ju’s political base and eliminating his powerful crony in the military such as Hur Bong-hak and Kim Kwang-hyop, Kim 100 Jong-il began to actively influence North Korean party politics in the late 1960s. Kim Jong-il had played a major role in appointing and dismissing of personnel as well as in leading political purges ever since he began his political career in the Worker’s Party in 1946. At the end, Kim Jong-il replaced his uncle as the party organization secretary in 1973 and solidified his status as heir in the following years. Kim Il-Sung: An Anti-Japanese Guerrilla, Who Rose to Power Through War, Purges and Personality Cult Kim Il-Sung is indeed the prototype of the first generation dictator who rose to power from humble background through military experience and force, and who ended up initiating war with the southern neighbor. He fought the Japanese expeditionary forces as a guerrilla during the Japanese occupation, which brought him political promi- nence after the Japanese withdrew from the Korean peninsular in 1945. After he created the Korean People’s Army with the support by the Soviet occupation authorities in 1948, he invaded South Korea to reunify the country militarily. Kim Il-sung thought the only way to reunify the country was by force of arms, and never advocated peaceful reuni- fication of Korea. After the war, when the various domestic factions challenged Kim’s political base, Kim Il-sung relentlessly purged those who fought alongside him as guer- rilla, let alone other political factions who operated in other parts of China and in the Soviet Union during the occupation. Kim Il-sung tightened his control and promoted himself as the true defender of the Korean revolution. By 1961, Kim managed to com- plete absolute power with no foreign occupation forces and no factions to challenges his position. Kim Il-sung was born on April 15, 1912 in Pyongyang to Kim Yong-jik and Kang Pan-sok, a peasant couple who struggled to provide the basic support to their three sons among whom Kim Il-sung was the eldest. Kim Il-sung’s formal education included elementary school and just two years of Chinese middle school in Manchuria 101 from which he was expelled at the eighth grade in 1929. His father died when Kim Il-sung was only fourteen, and his mother also died a few years later after Kim Il-sung joined guerilla bands. Thus, unlike the North Korean propaganda, Kim’s learning about communism was far from intellectual awakening through formal education, but was from random interactions with Korean Nationalists and Chinese Communist guerrillas during his years as a guerrilla (Suh 1995, 8). Kim Il-sung started to participate in the Chinese guerrilla groups against the Japanese forces after he was expelled from mid- dle school. He fought in eastern and southern Manchuria where the largest number of Korean partisans was concentrated, and Kim became one of the better-known Korean guerrilla leaders. In fact, by 1936, Kim Il-sung’s rise within the guerrilla was prominent, and his name became important enough for the Japanese to post a reward on his arrest. Numerous accounts of his activities such as the Liudaogou raid of April 26, 1938, and his raid into Korea in May 1939 confirm his accomplishments (Suh 1995, 37). In 1941 when the Japanese expeditionary campaign intensified, Kim Il-sung was the only guer- rilla leader who neither surrendered nor was captured or killed (Suh 1995, 35). North Koreans trace their revolutionary tradition to these partisan activities of Kim Il-sung in his anti-Japanese revolt, and his partisan group later came to dominate the core political leadership of North Korea. The Japanese expeditionary forces eventually defeated the guerrillas, and when Wei Zhengmin died on March 8, 1941, who was Kim’s superior and mentor, he fled to the Soviet Union. 87 While Kim Il-sung’s accomplishments during his anti-Japanese guerrilla years 87 David Sanger. (July 9, 1994). Kim Il Sung Dead At Age 82; Led North Korea for 5 Decades; Was Near Talks with South. The New York Times. Available at <http://www.nytimes.com/1994/07/09/world/ kim-il-sung-dead-at-age-82-led-north-korea-5-decades-was-near-talks-with-south. html?pagewanted=all> Accessed in January 2019. 102 had been exaggerated after he gained absolute power in North Korea and some of his own claims were preposterous insisting on nonexistent records and denying the contri- bution of others, his accomplishments are indeed many. He did persist in the hopeless fight against the powerful Japanese forces and scored some impressive victories. The Japanese posted reward for Kim Il-sung’s arrest for 20,000 yen in 1936, but the amount rose to 200,000 yen by 1939 (Suh 1995, 46). Many in his guerrilla band surrendered to the Japanese and then offered information to help arrest Kim, but he managed to survive and fled. Kim Il-sung stayed in the Soviet Union until 1945 when Korea was liberated from the Japanese forces. Korean partisans who survived in Manchuria eventually fled to the Soviet Union at some time between 1940 and 1941 and then were largely scattered in three Soviet training camps. Kim Il-sung was trained at Okeanskaya Field School (Suh 1995, 48). During his five-year stay in the Soviet Union, Kim Il-sung married Kim Jong-suk who had joined Kim’s guerrilla band in 1935. Kim Jong-suk gave birth to Kim Jong-il on February 16, 1942. When Japan was defeated and withdrew from the Korean Peninsula in 1945, Kim Il-sung returned to the northern part of Korea, and quickly was introduced by the Soviet occupation authorities as the head of the North Korean Provisional People’s Com- mittee in 1946. The Soviet Union had chosen Kim because he was trained in one of their training camps and wore their uniform. Once they chose Kim as the leader, they maneu- vered Kim Il-sung into power, sustained and supported him in the North. In retrospect, Kim Il-sung was lucky in that his association with the Soviet Union during the past five years was more than any Korean had ever had. With the support of the Soviets, Kim Il-sung created the Korean People’s Army. He consolidated his power with his guerrilla partisans who came to constitute the key members in the Korean People’s Army. In the end, in 1948 when Democratic People’s Republic of Korea was proclaimed in the North, 103 separate from the South, Kim was elected premier and approved his cabinet. In 1949, the official name of the Worker’s Party of Korea was adopted, and Kim was formally elected chairman. Less than five years after Kim Il-sung returned to Korea, he attempted to unify Korea militarily with his newly trained army. Kim Il-sung never thought a peaceful resolution to the divided Korea was an option, and he thought the only way to achieve national reunification was through use of force (Suh 1995, 112). Kim put all his efforts to build an army that was stronger than that in the South, and he assigned all his partisans to this effort. Kim had many units trained in Manchuria and brought combat troops from northeastern China into North Korea before he launched a full attack against the South in 1950. Kim Il-sung’s attempt to unify the country militarily led to a major international conflict involving the United Nations and many other countries siding with either the North or the South. However, his effort to militarily reunify with the South failed only to lead to a ceasefire agreed upon by China and the United States. Due to this failure, challenges to Kim’s power started to emerge. However, Kim Il-sung brilliantly used the failure in the Korean War to suppress his opposition and consolidate his power. Kim Il-sung faced challenges from three major groups domestically. The first group to challenge Kim was the “Domestic Group” that was made up of the local Com- munists who operated underground in Korea and Japan. The second group was the Yanan Group whose members were the returned revolutionaries from China. Lastly, there were the Soviet Koreans who returned to Korea with the Soviet army and stayed to participate in the political process of building the North Korean regime even after the Soviets withdrew. The Domestic Group emerged as a challenge during the Korean War when it became obvious that the war was going to be an unsuccessful conclusion. Pak Hon-yong was at the forefront of the effort to overthrow Kim and tried to establish him- self as the leader. Pak’s loyal supporters attempted a coup against Kim Il-sung in 1953 104 but failed. Everyone complicit in the coup was eventually purged and Pak Hon-yong was expelled from the party and eventually executed in the same year (Suh 1995, 130). Moreover, in order to eliminate domestic opposition forces, Kim carried out extensive purges in the name of punishing those who were cooperating with the Japanese. This policy had the effect of rooting out most of the local Communists who ultimately sub- mitted to the Japanese at one point or another under a lot of pressure during the Japanese occupation. The members of the Yanan Group were mostly purged by the time the Chi- nese troops withdrew from the North in 1958, and the Soviet-Koreans lost much when the Soviet Union withdrew did not fight for Korea during the Korean War. Ho Kai- I, for example, who was the highest-ranking Soviet-Korean was purged. By 1961, all domestic challenges have disappeared and only those loyal to Kim remained and were promoted into high positions. Kim Il-sung also succeeded in molding his past record and his partisan struggle into the only legitimate revolutionary tradition in Korea by then. It is often considered that by 1961 Kim Il-sung had succeeded in establishing his political power beyond challenge by any domestic factions. When building of Kim Il-sung’s absolute power was complete, one notable development was the rise of the military as an effect of the partisan promotion who were mostly generals on active duty. The promotion of the partisan generals to political prominence led to a series of daring and dangerous military adventures such as the seizing of the Pueblo and the Downing of the American EC-121 spy plane as well as the attempt to take the life of the South Korean president, Park Chung Hee, in the 1960s. These events were reflection of the overzealous and domineering partisan generals, and for Kim, their recklessness was taking him too far to the brink. Ultimately, some of the powerful partisan generals were also purged, and with the fall of those generals, Kim Il-sung established his absolute power beyond challenge. 105 In sum, Kim Il-sung’s experience in the anti-Japanese revolt as a guerrilla helped him solidify his political base after he returned to Korea when the Japanese with- drew. Being exposed to violent, bloody struggle over the course of his guerrilla activities in Manchuria, Kim il-sung had developed practical skills necessary to command mili- tary forces under his rule. He indeed had many units trained in Manchuria, and brought many combat troops to North Korea in preparation of war with the South. His partisans played a key role in the formation and operation of the Korean People’s Army, and Kim Il-sung was able to consolidate his political base in the newly established North Korean regime with his guerrilla group from Manchuria. His skills as a military commander, support from his partisans coupled with Kim’s firm conviction that the reunification of the country could be done only through military force drove Kim to initiate one of the major international wars in modern history. Kim Jong-Il: Spoiled But Ambitious Successor Works His Way Through Power Needless to say, Kim Jong-il’s ascension to power was largely enabled by his father’s absolute power as he himself consolidated his domestic base in North Korea before and after the Korean War. Right after graduation from Kim Il Sung University, Kim Jong-il was placed in the Central Committee of the Worker’s Party, the regime’s center of power, where he rose quickly ultimately becoming number two in the propa- ganda department. Afterwards, he left the Central Committee to work on the country’s cinema and opera where he, by all accounts, had put tireless efforts to remold them. During this time, Kim Jong-il saw successes such as the movie production in 1969 and the 1971 revolutionary opera Sea of Blood. Kim Jong-il also worked in Kim Il-sung’s military bodyguard organization where he served as major before he returned to the Central Committee of the party where he continued to take posts in the propaganda and agitation department. Such advancement in the regime’s nerve center, of course, would 106 not have been possible if not for his blood tie to the supreme leader of the country. In 1972 the regime removed the definition of “hereditary succession” in Dic- tionary of Political Terminologies which the earlier edition defined as “a reactionary custom of exploitative societies whereby certain positions or riches may be legally in- herited. Originally a product of slave societies, it was later adopted by feudal lords as a means to perpetuate dictatorial rule," (O 1990, 66) signaling that a relative of Kim Il-sung would succeed him. Following such moves, in 1973, the Central Committee elected Kim Jong-il to the party’s elite politburo and named him party secretary for or- ganization and guidance. Finally, an official announcement came in 1980 to the outside world that Kim Jong-il was on his track to succeed his father. It was not until Octo- ber 1980 at the Sixth Congress of the Workers Party of Korea that Kim Jong-il seemed to have removed all opposition to be publicly anointed as the successor. Kim Jong-il was named secretary of the party’s Secretariat right after his father, and was elected the fourth highest ranking member to the Politburo. He was the third highest ranking be- hind his father and O Jin-u in the Military Commission of the party. It is clear that Kim Jong-il’s takeover of the party was completed by then (Suh 1995, 281). North Korea’s version of propaganda emphasizes the “merit-based” succes- sion of the young Kim, and how Kim Il-sung refrained from doing anything that could be viewed as a plan to groom his son as the heir(Kim 1982). However, it is hard to give credit to the North Korea’s official account that Kim Jong-il’s ascension was purely based on merit. Not only were Kim Jong-il’s entire upbringing and life sheltered and influenced by the supreme power of the country’s god-like leader, but he was permit- ted to have his way on whatever issue he desired. Over the course of his career in the party, Kim Jong-il would not have had the same opportunities and advancements in the regime’s central department were it not for his father’s clout. During his first years in the party’s Central Committee, Kim Jong-il was permitted to ignore his responsibilities 107 and evade established rules and divisions of authority, something that would have been unthinkable for any ordinary novice. Kim Jong-il was quickly able to get an overview of the regime (Martin 2004, 236) as he flitted from one issue to another. He frequently accompanied his father on guidance tour traveling all corners of the country giving some guidance of his own, and in the process, he met people from all classes of the country. Moreover, Kim Jong-il didn’t refrain from clashing with senior officers (Martin 2004, 239). For example, during his post in the military bodyguard organization, the young Kim often clashed with the chief of the Bodyguard Bureau, O Baek-ryong, a former anti-Japanese guerrilla comrade of his father, who would have normally won the respect and obedience from a junior officer with no military experience. Despite the undeniably obvious birth right and the blood tie to the country’s supreme leader that enabled Kim Jong-il to wield power in every phase of his career, it is to be noted that there was little certainty that he would succeed his father’s absolute status as a future heir. Being a newly established regime, North Korea had no precedent for hereditary succession, nor was the regime well poised to claim one as a Communist country that customarily denounces hereditary succession as a regressive practice of the past. Even when succession by one of Kim Il-sung’s close relatives was widely speculated, Kim Jong-il was not considered to be the top contender for the post for a while. Instead, his once powerful uncle Kim Yong-ju was for a long time assumed to be the successor who had the backing of a number of influential people in the military as well as Kim Song-ae, Kim Jong-il’s stepmother. Another strong candidate for the post was Kim Pyong-il, Kim Jong-il’s stepbrother who had enjoyed a more affectionate, closer relationship with Kim Il-sung than Kim Jong-il ever had as a child. Moreover, Kim Jong-il faced several oppositions from the first-generation revolutionaries including top civilian officials such as Kim Dong-kyu, Kim Il, Pak Song-chol and a senior general O Jin-u who treated Kim Jong-il as a small child. Such competitions from his relatives 108 as well as political oppositions from the regime’s top officials meant for Kim Jong-il that he had to constantly prove himself as a capable, qualified heir and go after his political enemies to root out any oppositions to his succession. A number of South Korean scholars maintain that Kim Jong-il’s succession did not entirely owe to his birth right, but that he rather earned his father’s support and favor through demonstration of his leadership skills and utter devotion to the elder Kim (Jung 2014; SRLee 2015; Lee 2005). In sum, although Kim Jong-il’s tie to his father was undeniably the most sig- nificant factor behind Kim Jong-il’s rise to power, Kim Jong-il’s fate was neither pre- determined nor secured throughout his rise to power. On the one hand, Kim Jong-il had to compete against his relatives who also had laid their eyes on the next supreme leadership as well as political opponents who would rather back his rivals. To emerge as the victor, Kim Jong-il mastered the skills to win over potential political allies and root out perceived opponents from within the partisan group and the military. On the other hand, to win over his father’s favor, Kim Jong-il learned that he had to continue to show utter devotion to the elder Kim playing the role of the modest, filial son. Not only did he succeed in his role as a devoted son, but he also intensified the personality cult of Kim Il-sung that ultimately put Kim Il-sung on the highest pedestal throughout the country. Finally, Kim Jong-il did genuinely work tirelessly in the regime’s party to yield some successes especially in the country’s drama and opera. First, Kim Jong-il emerged as the winner in his competitions against his un- cle, Kim Yong-ju and his stepbrother, Kim Pyong-il, both of whom had affectionate and close ties with Kim Il-sung. Kim Il-sung is known to have deeply cared for his younger brother, Kim Yong-ju, who survived the period of Japanese hunt for the anti-Japanese partisans. After consolidating his power, Kim Il-sung appointed Kim Yong-ju as party secretary for organization and guidance in 1962, and by 1970 Yong-ju was elected to the 109 Politburo and ranked number six in the entire regime. Moreover, Yong-ju had a number of high-profile missions such as participating in the talks with South Korea during 1972- 173 as the negotiator against South Korean intelligence chief Lee Hu-rak. At that time Kim Yong-ju was widely viewed as likely successor to his brother (Martin 2004, 193). He had the support of many high officials as well as Kim Il-sung’s wife, Song-ae, who wanted to simply maximize the chance at the top post for her son, Pyong-il, expecting that Yong-ju’s tenure might be shorter than Jong-il’s due to his older age. A struggle be- tween Kim Jong-il and his uncle, Kim Yong-ju was formed right away when the younger Kim started his career in the regime’s party after graduating from college. Kim Yong- ju was Jong-il’s first official boss as director of the Central Committee’s Organization and Guidance Bureau, but clashes between these two men had Kim Il-sung transfer his young son to the party chapter in North Hamgyong Province. At the end, once a pre- sumed successor and a powerful man in the regime’s nerve center, Kim Yong-ju lost to his nephew, Kim Jong-il, who succeeded in winning over Kim Il-sung’s favor through utter devotion and development of the so-called “Kimilsungism” that rendered the Great Leader as an absolute deity. In 1973, Kim Yong-ju lost his post of party secretary for organization and guidance to Kim Jong-il who had also been elected to membership in the party’s elite politburo the same year. Since then, Kim Jong-il’s power outgrew his uncle’s and had far greater authority than Yong-ju had ever had. Ultimately selecting Kim Jong-il over Yong-ju, Kim Il-sung named Yong-ju vice-premier, a post that was quite negligible compared to what Yong-ju previously held. Kim Yong-ju eventually dropped from public view entirely after 1975 only to be shortly noted not until 1993. Consequently, his followers such as Ryu Jang-shik also fell out from public sight after 1975 (Martin 2004, 278). Another contender of Jong-il’s was his stepbrother, Kim Pyong-il, who was born in the early 1950s, and who is said to look like his father (do Kang 1995b).His 110 father was able to spend more time with Kim Pyong-il than he ever did with Kim Jong- il in the less frantic times after Pyong-il was born. After attending the North Korean military academy and learning to fly civilian aircraft in East Germany, Kim Pyong-il joined his father’s military bodyguard crops like Kim Jong-il did. Like Kim Jong-il, he rose quickly in the party soon being promoted to colonel and vice-head of the strategic department of the bodyguards. With Pyong-il’s mother, Song-ae, backing her son to succeed her husband’s top post, Kim Pyong-il was viewed as another strong contender for the job for a long time. However, in the competition against his stepbrother, Kim Jong-il again emerged as the winner. While patiently waiting for any chance to get Pyong-il in trouble, Kim Jong-il seized the moment when he caught Kim Pyong-il let his followers flatter him by saying “Long Live Kim Pyong-il.” This specific phrase was permitted only to Kim Il-sung only by then (do Kang 1995b).Kim Jong-il had been spying on Pyong-il through the party headquarters spy network, Room 10. Kim Jong- il reported his stepbrother’s alleged violation of the one-man rule to Kim Il-sung who became infuriated. He cut Pyong-il’s career in the party short and sent him overseas including Bulgaria and Finland as ambassador. Before leaving abroad, Pyong-il stayed in Pyongyang for a short period of time after his political career was dashed. Kim Jong- il during that time made sure that anyone who remained close to Pyong-il was pushed aside as well. Few people dared to maintain friendly relationship with Pyong-il out of fear of Kim Jong-il. Moreover, people also started to avoid Kim Yong-il, Pyong-il’s brother, a studious man who didn’t even seek a place in public life and was never a threat to Kim Jong-il. While Pyong-il ended up living abroad far away from the political center in North Korea, Yong-il lived an isolated life in Pyongyang mostly studying history. As for Kim Song-ae who was first supporting Jong-il’s uncle, Yong-ju, dreaming to see her own son succeed the top post one day, Kim Jong-il made sure that any clout carried by her on Kim Il-sung was diminished. He introduced two other women to Kim Il-sung 111 driving a wedge into his father’s marriage with Kim Song-ae. These women became Kim Il-sung’s favorite women and Jong-il’s plan to reduce his stepmother’s influence succeeded (do Kang 1995b). At the end, Kim Song-ae who was the chairwoman of the Korean Democratic Women’s Union and who ranked number 67 in the party in 1970 was found to have been demoted to number 105 in 1980 on the party Central Committee membership list (Martin 2004, 288). Aside from palace intrigues against potential successors, Kim Jong-il had sev- eral political opponents among the powerful first generation revolutionaries who had served as guerrilla protégé of Kim Il-sung in Manchuria. These people included Kim Dong-kyu, Kim Il, Pak Song-chol, and a senior general O Jin-u (Martin 2004, 277). They viewed Kim Jong-il as a young boy and were opposed to Kim Jong-il’s succes- sion. An extensive scale of purges of perceived rivals within the partisan group and the military began around 1967 when Kim Il-sung sought totalitarian rule based on one-man rule system. Kim Jong-il tirelessly led the harsh purges of high-ranking officials such as Kim To-man in the late 1960s in part as an effort to earn his father’s approval but also as an attempt to oust high-ranking supporters of his uncle. By the late 1970s, virtually all oppositions to Kim Jong-il was eliminated. To remove those opponents in the military and the partisan groups, Kim Jong- il mastered the skills of manipulation and forming factions at the cost of betraying some of his old close friends. For example, Kim Jong-il betrayed an old friend of his, Yi, who became at odds with O Jin-u, to get O Jin-u on his side in the competition against his uncle as well as eliminating his opponents in powerful posts (do Kang 1995b). Kim Jong-il ousted Yi from his powerful post as head of the overall political department and handed it over to O. In return, O proved to be useful in driving Kim Yong-ju out of power through unseating Yong-ju’s loyal military men such as Hur Bong-hak and Kim Kwang- yop. Moreover, Kim Jong-il didn’t hesitate to manipulate documents to emasculate his 112 opponents politically. Kim Dong-kyu of the party political bureau, who was the main opponent to Kim Jong-il’s succession was found to be manipulated by the young Kim so as to appear that he had been a traitor to the anti-Japanese movement. Kim Dong- kyu was eventually exiled from Pyongyang in 1977 and then sent to dissidents’ camp in Hwasong where he died in 1984. Other high officials who were opposed to Kim Jong-il’s succession faced similar fate: Kim Chang-bong, a minister, was sent to concentration camp along with his family where they died within a few years of malnutrition and diseases. Kim Jong-il learned from the fall of his stepbrother, Kim Pyong-il, the im- portance of retaining the role of the modest, loyal son to his father whose will, after all, mattered tremendously in the competition to succeeding his post. Kim Yong-ju un- derstood this as well, and thus, the competition between Kim Jong-il and Kim Yong-ju resulted in who succeeded in gaining Kim Il-sung’s favor. Kim Jong-il came out supe- rior in reading Kim Il-sung’s mind and showering him with flattery who was by then accustomed to adulation. For example, in wishing him a long life, Kim Jong-il estab- lished the Kim Il-sung Institute of Heath and Logenvity that would use human guinea pigs whose age and health conditions were similar to Kim Il-sung (Martin 2004, 195). Kim Jong-il also supervised the preparation for Kim Il-sung’s sixtieth birthday which turned out to be the most extravagant celebration. Later, Kim Jong-il coordinated the construction of the Arch of Triumph, the Tower of Juche, the Kim Il-ung Stadium, and a number of other monuments to commemorate his father’s seventieth birthday in 1982. At the heart of the rivalry between Kim Jong-il and Kim Yong-ju, however, lied the personality cult surrounding the Great Leader, and the struggle between the two men boiled down to who could put Kim Il-sung onto higher pedestal. The god-like, absolute authority of Kim Il-sung in North Korea as had been seen beginning in 1960s was largely a product of rivalry between Kim Jong-il and Kim Yong-ju who sought 113 Kim Il-sung’s favor and trust. Kim Jong-il, in particular, proved to be the master of promoting and establishing unthinking devotion to Kim Il-sung’s one-man rule based on a “monolithic” or “unitary” system, also called, “Kimilsungism”(Lim 2012), a term credited to Kim Jong-il. Kim Jong-il, for example, ordered to place busts of Kim Il- sung in every classroom all over the country, and had school children memorize Kim Il- sung, the Great Leader’s life story and thoughts and hymn the praises of Kim the father. It was under Kim Jong-il’s oversight and plan that Kim Il-sung transitioned to North Korea’s official deity, the father of the people. Kim Jong-il had introduced numerous formalities to demonstrate people’s absolute obedience to the Great Leader. Under Kim Jong-il’s order, people had to write and offer grand speeches swearing allegiance to the Great Leader, sing songs exalting him, and lay wreaths at Kim Il-sung’s statue. As Hwang Jang-yup described Kim Jong-il’s political ambition, the young Kim had from early on developed ambition for his own career and was determined to make his father’s power his own. 88 In the process, Kim Jong-il had demonstrated tremendous skills of manipulation and political infighting as well as reading his father’s psychology that ultimately contributed to the establishment of Kim Il-sung’s monolithic ideological system (Lim 2012). Lastly, Kim Jong-il tried to prove his qualifications through hard work and success in some fields of the country, especially in the country’s cinema and opera. Kim Jong-il did genuinely work hard to remold North Korea’s filmmaking. His production of the Sea of Blood in 1971 was by all accounts a huge success that gave Kim Jong-il his much needed afterglow in the preparation of succeeding his father. The production helped Kim Jong-il to solidify his power and made him a strong, qualified successor 88 Hwang Jang-yop, Problems of Human Rights in North Korea. Available at <http://www2.law.columbia.edu/course_00S_L9436_001/North%20Korea% 20materials/hwang%20jang3.html> Accessed Through January 2019. 114 for his father’s job (do Kang 1995a). In general, Kim Jong-il’s work ethic was known to be quite rigorous, and his governing style did not exactly take that of a lazy man’s. According to Hwang Jang-yup, Kim Jong-il established a system of having the govern- ment department submit policy recommendation for his approval. No matter how busy he was, Kim Jong-il would read all the policy recommendations and provide his com- ments, place them in his special envelope with the recipient’s name on it, and personally seal it (Martin 2004, 284). This all amounted to a huge workload, but the young Kim never passed it to anyone else to handle it. In sum, while it is indisputable that Kim Jong-il’s rise to power was owed to his father’s clout, he had to fight to earn the elder Kim’s trust and favor amidst the competitions against his uncle as well as his stepbrother, both of whom had Kim Il- sung’s deep affection. Moreover, with no precedence of hereditary succession, Kim Jong-il faced numerous oppositions from high-ranking officials in powerful groups who had Kim Il-sung’s trust and affection. In his ascension to power, Kim Jong-il did have to go through various testing processes and palace intrigues that a typical heir in an established hereditary regime would not have to. In other words, Kim Jong-il did select himself into power with the very strong ambition and aspiration to make his father’s absolute status his own. Mechanism 2: Shared Narcissism and Self-Importance: Dispositional Traits and Foreign Policy The second mechanism revolves around leader’s dispositions and personality traits that help shape his foreign policy decisions. Power Seizers tend to have greater am- bitions and risk tolerance than a typical leader, which contributed to their rise to power in the first place. Moreover, building on previous research that maintains personalist dictators have the insatiable desire for attention and dominance over others (Glad 2002), 115 I argue that the Power Seizers are more willing to take upon conflictual and risky path. Kim Il-sung indeed is the prototype of a Power Seizer who rose to power via force and eventually took a military path to fulfill his ambition to lead a unified country. Despite the setback from the unsuccessful war operation during the Korean War, he managed to attain a position within North Korea that fulfilled his personal ambition and self-ego. On the other hand, my argument goes that the Power Successors may lack the desire for dominance over others and the ambition for status and power based on the as- sumption that the typical Power Successors didn’t have to seek the power in their hands as the Power Seizers did. If, however, the Power Successor did have to fight for the status and power, we could expect that he may display some of the same dispositional traits that the Power Seizer possesses. As previously discussed, Kim Jong-il’s rise to power, although highly owed to his ties to the Great Leader, was a tumultuous one in that he had to work through a lot of competitions against close relatives and political op- positions by powerful high-ranking protégés of his father in a country where hereditary succession had not yet been the norm. Like his father, Kim Jong-il was ambitious and had an enormous self-regard. He was determined to make his father’s status and power his own and took every measure to climb to the top position without a slip. In sum, while both Kim Il-sung and Kim Jong-il share the same characteristics as being ambitious and narcissistic. These were important qualities that kept them moving forward to the top position. Kim Il-sung: Fatherless and Uneducated, But Ambitious and Self-Absorbed Kim Il-sung was born on April 15, 1912 at Mangyongdae near Pyongyang, a place later named as the “cradle of the world revolution.” 89 Kim Il-sung later portrayed 89 July 11, 1994. Kim Il Sung. The Telegraph. Available at <http://www.telegraph.co.uk/news/obituaries/7859377/Kim-Il-Sung.html > Accessed 116 not only himself but his entire family as the people who were destined to fight for the Korean revolution against the greater foreign powers around the Peninsula. He main- tained that his grandfather set fire to the American “pirate ship”, The General Sherman, while sailing up the Taedong River to Pyongyang in 1866. 90 And his father is portrayed to have faced death by Japanese torture. 91 In fact, his father ran a herb pharmacy and died when Kim Il-sung was 14, and his mother died two years later. These stories, most of them unverified, seemed to be an effort on the part of Kim to project himself as a pious son. However, his show of filial piety does not seem to stem from true reverence of his parents, who rather did little for Kim Il-sung during his boyhood (?, 53) Rather, Kim Il-sung seemed to have been keen on building his own image as a pious son from a revolutionary Korean family, and thus, the purpose was entirely self-serving (Suh 1995, 53). As most, if not all, observers of North Korea and Kim Il-sung maintain, the major impediment to reading Kim Il-sung’ true character are the extravagant and unsub- stantiated accounts like these that he and his followers had put forth over the decades after Kim Il-sung consolidated his base. However, that very problem reveals some im- portant qualities of his personality: his enormous self-regard and ambition. in January 2019. 90 David Sanger. (July 9, 1994). Kim Il Sung Dead At Age 82; Led North Korea for 5 Decades; Was Near Talks with South. The New York Times. Available at <http://www.nytimes.com/1994/07/09/world/kim-il-sung-dead-at-age-82- led-north-korea-5-decades-was-near-talks-with-south.html?pagewanted=all > Accessed in January 2019. 91 July 11, 1994. Kim Il Sung. The Telegraph. Available at <http://www.telegraph.co.uk/news/obituaries/7859377/Kim-Il-Sung.html > Accessed in January 2019. 117 Kim himself wrote after his personality cult intensified that he had been the center of the Korean resistance movement as early as 1930s and that the enemy was convinced that the Korean army against the Japanese and Manchukuko would have col- lapsed were it not for him (Kim 1995). Kim Il-sung indeed developed as a young man a penchant for the company of people who regarded him as a great man, and hero (Martin 2004, 46). He relished praises by the people and was never tired of sycophants around him. At the end of his four-decade rule in North Korea, it is fair to say that his person- ality cult in North Korea operated as a religion. People sobbed at the benevolence and mercifulness of the Great Leader, and accounts abound as to his kindness granted upon individual North Koreans. For his sixtieth birthday, Kim Il-sung built the most expen- sive mansion, the largest monuments in North Korea, and was greeted with elaborate fanfare at the most extravagant party. Due to Kim Il-sung’s narcissism and his followers’, especially his own son, Kim Jong-il’s, efforts to reinforce it, North Korea operates not as a communist state but a peculiar brand of oriental despotism (Martin 2004, 315) where the Great Leader, Kim Il-sung is the merciful father of all North Koreans who were taught and brainwashed from their birth to put Kim Il-sung before their own parents. Kim Il-sung was called many things that no dictator had ever been called: the great leader, national hero, ever- victorious iron-willed brilliant commander, an outstanding leader of the international Communist movement, an ingenious thinker, the sun of the nation, the red sun of the oppressed people of the world, the greatest leader of our time (Martin 2004, 316). Kim Il-sung was fascinated by praise and admiration by the public, and Kim Jong-il who played the role of a pious son enthusiastically created a personality cult for his father. Kim Il-sung wrote in his memoirs that “A man who enjoys the love of the people is happy, and a man who does not is unhappy. This is the view of the nature of happiness which I have maintained throughout my life”(Kim 1995). So what he built 118 in North Korea was a personal kingdom where people were his loyal subjects rather than a communist state. In the end, Kim Il-sung managed to rewrite the constitution in 1972 and had himself elected president of the republic as well as chairman of the party (Martin 2004, 251). As self-absorbed as he was, Kim Il-sung does deserve to be evaluated on his own terms whose dramatic rise and consolidation of absolute power across an entire nation reflects perseverance and ambition of a young man from humble background. He persisted as guerrilla fighter when most other Koreans in his situation eventually succumbed to the Japanese forces if they were not killed. As an uneducated young man expelled from Middle School, he joined the guerilla band surviving and thriving during his exile. He was brave and tough and never gave in (Martin 2004, 46). Biographer Suh gives credit to Kim Il-sung in that regard saying that “it is his persistence and obstinate will, characteristics of many successful revolutionaries, that deserve recognition” (Suh 1995, 54). Martin observes that “Kim comes across as a young man who matured fast to become an effective and shrewd leader. . . . [H]e showed considerable intelligence and common sense”(Martin 2004, 46). When Kim Il-sung fled to the Soviet Union, he managed to impress his Soviet mentors with his political shrewdness despite his physical shortcomings. Yu Song-chol, who was Kim Il-sung’s Russian interpreter during Kim’s stay in the Soviet Union re- called him being “obedient and ardent toward Brigade Commander and the Soviet of- ficers” (Martin 2004, 49). Kim Il-sung, the fatherless young revolutionary, seemed to have the ability to appeal to paternal instincts of elders and superiors from his middle school days in Manchuria who came to support him and help him on his way up and forward. Kim Il-sung had to rely on the support of his Soviet mentors and knew how to flatter them carrying out their instructions as given, who, in turn, would reward Kim with more power and autonomy (Martin 2004, 49). 119 At the same time, Kim Il-sung was not physically the strongest in the brigade. He was “lean and weak, and his mouth always open, perhaps due to blockage in his nasal passages” (Martin 2004, 49). According to Yu Song-chol, Kim Il-sung became so exhausted during ski training that he had to rely on his subordinate for support in order to move (Martin 2004, 49). However, such physical weaknesses didn’t seem to pose any problem for Kim Il-sung to stand out to the Soviet officers. The Soviets liked Kim Il-sung finding him “exceptionally smart and possessive of leadership qualities”(Martin 2004, 49). On the other hand, Yu recalled that Kim Il-sung expected absolute obedi- ence and was stern and cold, unemotional and calm to his subordinates although that depended on whether the subordinate served alongside him in the guerrilla band during his days in Manchuria (Martin 2004, 49). The Korean war that was planned by Kim Il-sung a few years after he ascended to top leadership in North Korea marked the epitome of Kim’s ambition to unify the country militarily. Convinced that he could take the South within three days, Kim Il- sung brilliantly played Stalin and Mao against each other, who were first both skeptical and fearful of a potential American intervention. 92 Kim skillfully manipulated his far bigger partners by inflating Mao’s interest when talking to Stalin, and Stalin’s when reaching out to Mao. In the end, Kim’s ambition led to an outbreak of a total war that involved the intervention of the United Nations and major countries such as the United States and China causing 1,000,000 civilian deaths in the South and 600,000 in the North, with 250,000 military deaths from South Korea and the United States 92 David Sanger. (July 9, 1994). Kim Il Sung Dead At Age 82; Led North Korea for 5 Decades; Was Near Talks with South. The New York Times. Available at <http://www.nytimes.com/1994/07/09/world/ kim-il-sung-dead-at-age-82-led-north-korea-5-decades-was-near-talks-with-south. html?pagewanted=all> Accessed in January 2019. 120 combined. 93 The war, however, did not go as Kim planned and ended with an armistice agreed by the United States and China. Despite the unsuccessful war operation and the worsened economy in the North, Kim Il-sung managed to not only survive politically but used the failure as an opportunity to wipe out his remaining rivals to further strengthen his domestic position. In the end, Kim Il-sung for all the damage he had done to the country and the people, he demonstrated mastery of manipulative skills that he used to satisfy his large self-ego and ambition to power. Kim Jong-il: Spoiled and Ambitious Young Man Seeks Absolute Power Kim Jong-il had a very narrow and sheltered upbringing. He was born on February 16, 1942, in the Soviet Union but returned to Korea when he was three. Kim Jong-il was the eldest of three children, two boys and one girl. His brother died in 1947 and lost his mother Kim Jong-suk in 1949. During the Korean War, he spent most of the years in China, surrounded by relatives and North Korean elites. Accordingly, he was subject to little interaction that would challenge him as the son of the North Korean absolute dictator. Being surrounded by members of the North Korean elites and sycophants, he was encouraged to have his own way and faced no obstacles in his path. In particular, no one had control over the young Kim since his mother, Kim Jong-suk, died in 1949. Kim Jong-il behaved like a prince always permitted to have his way. His conceit and narcissism was coupled with his growing ambition to make his father’s power and status his own. Kim Jong-il knew that his father’s absolute power wasn’t ultimately his without the effort to being appointed as the successor. In short, Kim 93 Korean War Fast Facts (Updated on April 30, 2018). CNN. Available at <http://www.cnn.com/2013/06/28/world/asia/korean-war-fast-facts/index.html> Ac- cessed Through January 2019. 121 Jong-il possessed the qualities that helped him ascend to the throne: Narcissism and ambition. Kim Jong-il from early on had developed a reputation of being “wild, reck- lessly impulsive and, by turns, cruel and warm-hearted, even extravagantly generous.” (Martin 2004, 203). During his teen, the young Kim, rather than preferring to study, liked to watch movies with fellow students. Kang Yong-gu, a former superintendent of Namsan Junior High School and Kim Dan, a classmate of Kim Jong-il’s at Namsan Elementary School in Pyongyang testified that Kim Jong-il was “self-centered and his behavior was impolite”(Martin 2004, 212).. Kim Jong-il was said to boast his expensive belongings to his friends and always drove expensive cars or motorbikes at a fast speed. He liked to savor attention and receive gifts from his followers such as gold bullion and foreign currency. In turn, Kim Jong-il also tended to give gifts lavishly. Several anecdotes by North Korean defectors who knew Kim Jong-il confirm his extravagance. A former official, Kang Myong-do testified that Kim Jong-il gave gifts to buy people off and bring them over to his side. He would give expensive, imported cars with the license plate ‘216’ which represent the numerals of his birthday, February 16 (Martin 2004, 276). Kim Jong-il’s self-absorption and narcissism are accompanied by his jealousy about other people’s popularity and their good fortune. He would become jealous about his subordinates in the party if they were popular among the masses, and would soon replace them with other officials. Furthermore, he would even purge those officials denouncing their “anti-revolutionary” mindset for being a distraction from the one-man rule system towards the Great Leader. 94 Kim Jong-il justified his purges of those popular 94 Hwang Jang-yop, Problems of Human Rights in North Korea. Available at <http://www2.law.columbia.edu/course_00S_L9436_001/North%20Korea% 20materials/hwang%20jang3.html> Accessed Through January 2019. 122 officials using ideological terms that only the Great Leader is deserving of worship by the masses, and any ordinary party official should oppose adoration among the people directed to them. 95 In the same line, Kim Jong-il maintains that people should attend to the Great Leader only and should oppose any close relationship or social circles that does not revolve around the Great Leader. He is against family orientations, bonds based on regional origin, and opposes teacher-student or senior-junior relationships and all forms of social gathering other than his own. Kim Jong-il tended to frequently hold drinking parties as a means of forming his own faction and keeping other gatherings excluding himself at bay. 96 He used his drinking parties to closely watch his subordinates and people around him and imbue them with a sense of pride at being close to the Great Leader. Kim Jong-il’s entirely egotistic and narcissistic viewpoint of ideology and sta- tus grew into an ambition and determination to inherit his father’s power. Kim Jong-il was a conceited boy. 97 The young Kim always had his friends play the role of ‘minis- ters’ while giving them orders. Later on, he tended to dominate meetings and seminars to conclude all discussions congruent with his opinion. 98 He flaunted his status as a young boy and developed a strong ambition to make his father’s absolute power his own. Hwang Jang-yop noted that Kim Jong-il was “more aggressive than his father in 95 Ibid. 96 Ibid. 97 Hwang Jang-yop, Testimonies of North Korean Defectors: True Picture of North Korea According to a Former Workers’ Party Secretary Available at <https://fas.org/irp/world/rok/nis-docs/hwang2.htm > Accessed Through January 2019. 98 Ibid. 123 the process of power inheritance.” 99 His ambition for absolute power meant lack of trust in anyone in the palace and in the party. Kim Jong-il had great skills of manipulation and was used to political infighting being exposed to high statecraft and palace intrigue on a daily basis. He had trusted no one but relied on a system of surveillance of virtually every officials in North Korea, high and low-ranking. After Kim Jong-il became in charge of the Central Com- mittee, the regime’s nerve center, surveillance of even the top-ranking officials became intense. 100 Kim Jong-il set up a new committee designed to specifically control the top officials in the party Central Committee carrying out secret surveillance and intelligence activities. The officials in the party were controlled under multiple fold scrutiny at all times under Kim Jong-il. Moreover, the Three Revolution movement, which was first created under Kim Il-sung to purportedly function as morale-building, worked as snoops and spies under Kim Jong-il’s command. The Three Revolutions were organized teams of twenty to fifty young party members, intellectuals and college students who were sent into various sites including factories, farms and schools to control potentially “trouble- some” members in the society. Hwang Jang-yop who had sustained close involvement with Kim Jong-il as Chairman of the Standing Committee recalled Kim Jong-il as someone who “does not like living in harmony with others.” 101 He is by nature conflict-prone and had a great aptitude for dictatorship. He does not like people to trust each other but want them to rely only on him. For example, at the Central Committee meetings, instead of empha- sizing positive developments, Kim Jong-il would encourage mutual criticism among the 99 Ibid., Lee (2005) 100 Ibid. 101 Ibid. 124 officials and ask them to point out the negative. Participants who were hesitant to crit- icize others were blamed as lacking revolutionary zeal and loyalty to the Great Leader. Only those who harshly criticize others were deemed to be loyal to Kim Jong-il during the mutual denunciation sessions. According to Hwang, Kim Jong-il would sit back and watch as the officials fight with each other. Not only was Kim Jong-il conflict-prone but he could also be very cruel. As previously discussed, he would not hesitate to sacrifice his close friends for political gain. In pursuit of his political ambition, Kim Jong-il had cruelly killed countless peo- ple. 102 One anecdote told by Hwang goes that he had an official kill his own wife during a party for having written to the Great Leader about the young Kim’s life of debauchery. In sum, Kim Jong-il was both conceited and ambitious. He flaunted his status and power as the son of the Great Leader of the country, but he also knew that he had to work his way through to the post his father held and was ambitious enough to ultimately succeed to make it his own. In the succession process, he demonstrated mastery over manipulation and infighting that was cruel and violent. Self-importance and ambition are two common qualities that belong to most dictators, and they characterize Kim Jong- il as well. 102 Ibid. 125 Chapter 9. Conclusion This dissertation attempts to provide a more comprehensive analysis on auto- crats’ conflict behavior by offering an alternative criterion: self-selection into power. I argue that the Power Seizers who selected themselves into power tend to initiate more international conflict than the Power Successors who did not. The conventional two- dimension typology based on personalism and militarism conflates the leaders’ individ- ual variation and the extent of institutional constraints making it hard to disintegrate its own independent effect from the other. Existing theories derived from this typol- ogy raise the question of whether leaders in personalist regimes are typically belligerent because of the nature of the political institutions he rules in or because of the leader’s peculiarities that helped create weak domestic institution in the first place. Similarly, we are unsure as to whether military leaders are particularly hostile because of his own personal penchant for use of force or because of the domestic military institution that encourages tough behavior. This dissertation introduces a new framework that controls for the institutional effect on autocrats’ initiation of international conflict and shows that leaders’ predisposition to use of force is statistically and substantively significant. This dissertation not only helps us better understand the separate effect of leader disposition devoid of institutional influence, but also introduces us to new types of autocrats IR scholars have long ignored. The hereditary Power Seizers generally refer to the Arab monarchies whose omission in the sample of dictatorship with regards to foreign policy behavior has almost been a standard practice in IR scholarship. The inclusion of the nonconstitutional Power Successors into the broader subset of hereditary regimes(Brownlee 2007) encourages a more systematic analysis of autocrats based on the common criteria of self-selection. Holding against this single, common criterion, I collect a sample of Power Successors that go beyond the Arab region and beyond the 126 traditional monarchs to include some atypical regimes for the sample, such as Taiwan and Singapore. The findings suggest that my argument holds across the different regions and regime types. Finally, this dissertation combines insights from both the IR on conflict stud- ies and comparative politics on authoritarianism, and attempts to bridge the missing connection between hereditary regimes and their foreign policy behavior. My findings suggest that the monarchs in the Middle East are not only remarkably resilient spared of most of domestic turmoil in the past decades, but they are also relatively peaceful agents in dealing with international affairs. If so, contrary to earlier scholarly verdict that hereditary rule “has come to an end” (Burling 1974), monarchies have proven to be a viable form of an effective negotiator internationally. Furthermore, the paramount domestic position of power by kings, emirs, and sultans has us rather predict impulsive and reckless foreign policy behavior. Yet, the pattern of relative nonviolence amongst these leaders against other international agents compels us to rethink the relationship between domestic power and foreign policy behavior. In short, not every leader pursues aggressive foreign policy even if they are free to set any goal at their whim and are unconstrained by concerns of being punished at home. The findings also cast a number of policy implications for analysts and diplo- mats facing autocrats abroad. The Power Successors are largely peaceful and cautious when it comes to international conflict. Previous research has shown that monarchs are unlikely to fall, and if they do, most likely to fall in the hands of violent dicta- tors(Geddes, Wright and Frantz 2014). In the post 1945-period, only one out of eight countries in which monarchical rule ended has been replaced by democracy, and most of them were taken control by some of the most repressive dictators, such as Qaddafi replacing the Libyan monarch, Khomeini the Iranian monarch, and Mubarak the Egyp- tian. Moreover, the years after the overthrow of monarchs in Yemen, Ethiopia, and 127 Afghanistan have seen long, bloody civil war in all these countries. Indeed, given the stability and peace by the Power Successors, and the nonconstitutional Power Succes- sors in particular, monarchical rules may be preferable on international security grounds than autocrats that come after replacing monarchs. If a monarchic regime should fall, foreign policy experts should pay careful attention to the leadership that seizes power. History has shown that the emergence of violent Power Seizers such as Hosni Mubarak, Ben Ali, and Qaddafi came in the aftermath of overthrow of monarchs. The findings raise another important issue surrounding the Power Seizers and Power Successors: what can we infer about the extent of domestic political liberaliza- tion initiatives and repression of domestic opponents in those regimes? If we follow the same logic outlined through the framework presented in this dissertation, we would expect that domestic repression is more severe under the Power Seizers than under the Power Successors. Previous research has found that Middle Eastern monarchs are more conducive to the rule of law and less corruption, and that they are not systematically as- sociated with civil protest (Menaldo, 2012; Spinks, Sahliyeh, and Calfano, 2008) com- pared to Middle Eastern republics. However, it is evident that domestic repression is a pressing issue in some regimes under the Power Successors. It is a well-documented fact that, for example, in North Korea and Syria, imprisonment and execution of political dissidents are a common practice. As of today, an estimated 80,000 to 120,000 political prisoners are held in detention camps in North Korea, where inmates face abuse, torture, starvation rations, and forced labor. 103 While plagued by civil war, the Assad regime in Syria maintains complete dominance over intelligence and security apparatus to monitor 103 Freedom In the World. Freedom House. https://freedomhouse.org/report/ freedom-world/2018/north-korea (Accessed Through January, 2019) 128 and punish opposition political activities. 104 The stories of those regimes under Power Successors cast a gloomy prospect that the domestic suppression and physical repres- sion may continue as long as their obsession to hold on to supreme power continues. If international conflict should arise by these Power Successors, it may be more about enhancing their domestic base and distracting attention from domestic challenges than some greater desire for international expansion per se which would be the case for the Power Seizers. Finally, the Power Successors might appear as small in absolute number, but like the Power Seizers, they are likely to form and emerge at any time and at any event. To the extent that autocrats prevail, there is always a chance that the world will see an- other Power Successor. Recall that Qaddafi was grooming his third son for succession, and so was Mubarak in Egypt before these regimes were swept by mass demonstrations during the Arab Spring. Once established, however, those Power Successors in “noncon- stitutional” dynasties are likely to inherit more tumultuous domestic political landscape than their counterparts in the constitutional monarchs, as evidenced by the fact that most “nonconstitutional” Power Successors have been unable to pass down power beyond the third generation yet. This suggests that even if less violent internationally, the Power Successors might be highly likely to be replaced by domestic political opponents. Thus, foreign policy analysts should take heed of the potentially volcanic domestic conditions harbored in those regimes run by nonconstitutional Power Successors. 104 Freedom In the World. Freedom House. https://freedomhouse.org/report/ freedom-world/2018/syria (Accessed Through January 2019.) 129 Appendix: Additional Hypotheses Testing Collapsing Constitutional and Nonconstitutional Power Successors: All Power Successors vs. Power Seizers The distinctions between the constitutional monarchies such as Saudi Arabia and Kuwait and the nonconstitutional dynasties such as North Korea and Syria are made because of the theoretical expectation that the nonconstitutional dynasties are more belligerent than constitutional monarchies. However, if we were to take the most overarching criteria of “self-selection into power by force” only, we could collapse the constitutional and non- constitutional Power Successors, and expect that all Power Successors are less conflict prone than the Power Seizers. A.H1: (All) Power Successors are less likely to initiate international conflict than Power Seizers. I collapse the constitutional and the nonconstitutional leaders in the Power Successors to compare the rates of conflict initiation based on the “self-selection into power by force” criteria only. I predict that the coefficient estimates for the Power Seizers to be bigger than those of the Power Successors, and that their differences are statistically significant through Wald Test. The results from Table 10 strongly support my hypothesis both statistically and substantively. Power Seizers are more likely to initiate MIDs compared to Power Successors by more than 100% in rate across all models, and the rates for the fixed effects models are even stronger indicating an increase of 164% in rate for model (2) 130 and 173% for model (4). 105 Wald Tests indicate that these differences are statistically significant at p<0.0000 for all four models. 105 An increase of 134% in rate for model (1); 164% for model (2); 137% for model (3); 173% for model (4). 131 Table 10: MID Initiation by the Power Seizers vs. Collapsed Power Successors (1) (2) FE (3) (4) FE Any MIDs Any MIDs Higher Levels of MIDs Higher Levels of MIDs Power Seizers 0.317 a 0.186 b 0.345 c 0.218 d (0.087) (0.073) (0.088) (0.074) Power Successors -0.534 a -0.786 b -0.520 c -0.789 d (0.172) (0.186) (0.170) (0.190) Trade Openness -0.344 -0.104 -0.356 -0.058 (0.147) (0.091) (0.157) (0.103) Years since Regime Est. 0.002 0.003 0.002 0.004 (0.003) (0.002) (0.003) (0.002) Capability 2.372 -0.349 2.401 -0.214 (1.110) (1.262) (1.106) (1.291) Total Borders 0.067 0.066 (0.017) (0.017) Major Power Status 0.418 14.076 0.428 15.003 (0.133) (421.052) (0.130) (546.439) Total Allies 0.011 0.005 0.014 0.012 (0.017) (0.016) (0.017) (0.016) Years Since Last MID -0.022 0.041 -0.021 0.041 (0.010) (0.013) (0.010) (0.013) Cubic Splines YES YES YES YES Region Dummies YES NO YES NO Constant -0.950 -0.927 (0.342) (0.346) Observations 7540 6922 7540 6874 Negative binomial regressions with standard errors in parentheses clustered by country. t p< 0.10, p< 0.05, p< .01, p< .001 Models (2) and (4) are fixed effects of models (1) and (3), respectively, and thus drop time-invariant variables such as contiguity (Total Borders) and region dummies. a Wald Test: different at p< 0.0000. b Wald Test: different at p< 0.0000. c Wald Test: different at p< 0.0000. d Wald Test: different at p< 0.0000. 132 Hypothesis: Power Successors vs. Democracies A.H2: (All) Power Successors are no more likely to initiate conflicts than democracies. Using the collapsed Power Successors as the main IV , and Democracies as the base category in the analyses, I hypothesize that the Power Successors are no more likely to initiate conflicts than democracies. The results from Table 11 strongly support my hypothesis both statistically and substantively. Power Successors are less likely to initiate international conflicts than democracies by as high as 52.1% in rate, and by as low as 38.2% in rate, 106 and the results are highly statistically significant at either p<0.01 or p<0.001. 106 A decrease of 39.1% in rate for model (1); 52.1% for model (2); 38.2% for model (3); 51.9% for model (4). 133 Table 11: MID Initiation by the Collapsed Power Successors Compared to Democ- racies (1) (2) FE (3) (4) FE Any MIDs Any MIDs Higher Levels of MIDs Higher Levels of MIDs Power Seizers 0.372 0.248 0.401 0.286 (0.108) (0.100) (0.109) (0.103) Power Successors -0.496 -0.736 -0.481 -0.732 (0.178) (0.199) (0.178) (0.205) Other Autocrats 0.107 0.118 0.110 0.125 (0.107) (0.099) (0.105) (0.102) Trade Openness -0.336 -0.096 -0.348 -0.048 (0.145) (0.089) (0.155) (0.102) Years since Regime Est. 0.003 0.004 0.003 0.004 (0.003) (0.002) (0.003) (0.002) Capability 2.280 -0.532 2.308 -0.408 (1.156) (1.270) (1.164) (1.300) Total Borders 0.066 0.065 (0.016) (0.017) Major Power Status 0.425 14.042 0.435 22.530 (0.144) (441.753) (0.140) (.) Total Allies 0.009 0.006 0.012 0.012 (0.017) (0.016) (0.017) (0.016) Years Since Last MID -0.022 0.041 -0.022 0.042 (0.010) (0.013) (0.010) (0.013) Cubic Splines YES YES YES YES Region Dummies YES NO YES NO Constant -0.949 -0.926 (0.342) (0.345) Observations 7490 6884 7490 6836 Negative binomial regressions with standard errors in parentheses clustered by country. t p< 0.10, p< 0.05, p< .01, p< .001 Models (2) and (4) are fixed effects of models of (1) and (3), respectively, and thus drop time-invariant variables such as contiguity (Total Borders) and region dummies. 134 Distinguishing Power Seizers and Comparing with Leaders Under Similar Constraints : Power Seizers (hereditary) vs. Similar Con- straints; Power Seizers (nonhereditary) vs. Similar Constraints A.H3: Power Seizers are more likely to initiate international conflict than other autocrats under similar constraints even when distinguished between the "heredi- tary" and "nonhereditary" Power Seizers. I compare the coefficients of Power Seizers with those for Similar Constraints, carrying out two-tailed Wald Tests to evaluate the significance of the differences in these coefficients (Table 12). Next, I exponentiate the coefficients to interpret the effects of predictor variables in negative binomial regressions by recovering the difference in the logs of expected counts of MIDs for a one unit change. 107 For example, the coefficient of 0.314 for Power Seizers (nonhereditary) in model (1) (Table 12) indicates that the rate for MID would be expected to increase by a factor of 1.369, 108 or 36.9% of rate, if the leader is a Power Seizers (nonhereditary), while holding all other variables constant. I expect that the coefficient estimates for Power Seizers to be larger than those of Similar Constraints, and more importantly, that the differences between these coefficients to prove statistically significant through Wald Test. I do not expect, however, that the coefficient estimates for Similar Constraints to be necessarily negative, nor those for Power Seizers to be positive. The results from Table 12 indicate that MIDs are more likely to occur by as 107 The coefficients in negative binomial regressions are usually reported in terms of incidence rate ratios, which can be written as b = log(m x 0 +1 ) log(m x 0 ) = log(m x 0 +1 /m x 0 ), where b is the regression coefficient,m is the expected count. 108 The exponent of 0.314 is 1.369. 135 low as 36.9% and as high as 43.1% 109 in rate if the nonhereditary Power Seizers are in power compared to when they are not. These effects are highly statistically significant across all models. Furthermore, the Wald Tests indicate that the differences between the regimes under the Power Seizers and regimes under Similar Constraints are statis- tically significant at either p<0.0004 or p<0.0001. Substantively, the differences are as striking. There is an increased rate that MIDs will occur by Power Seizers (non- hereditary) than by Similar Constraints (nonhereditary) by a percentage rate ranging from 39.1% to 49.4%. 110 Thus, the evidence supports my hypothesis that nonheredi- tary Power Seizers are more likely to initiate conflict than other regimes with similar institutional constraints on the executives. Figure 9 illustrates the substantive effects of all three previously tested hy- potheses using CLARIFY (Tomz et al. 2003). I estimate the expected probability of MID initiation when all of the covariates are set to the observed values for Iraq in 1990. 111 The figure indicates that, had Iraq been ruled by a dictator other than Saddam Hussein in 1990, there would have been a 12 % decrease (57.52-45.99, column 1, row 1)in chance that Iraq would invade Kuwait despite similar institutional constraints, and a decrease of 28 % (61.34-33.37, column 1, row 3) if he were a Power Successor(constitutional). The point estimates are in the same expected directions, although as discussed earlier, 109 36.9% for model (1) ; 37% for model (2); 42.3% for model (3); and 43.1% for model (4). 110 The % change is calculated using the formula:(e bb 0 1)100, where b is the coefficient estimate of Power Seizers (nonhereditary) and b’ is the coefficient estimate of Similar Constraints (nonhereditary). An increased rate of 48.5% for model (1); 39.1% for model (2); 49.4% for model (3); and 41% for model (4). 111 The figures for the first column are based off of model (1) of Tables 12, 6, and 7, and those for second column are of model (4) of Tables 12, 6, and 7. 136 Table 12: MID Initiation by Power Seizers (when distinguished between "heredi- tary" and "nonhereditary") vs. Leaders Similar Institutional Constraints (1) (2) FE (3) (4) FE (5) (6) FE (7) (8) FE Any MIDs Any MIDs Higher Levels of MIDs Higher Levels of MIDs Any MIDs Any MIDs Higher Levels of MIDs Higher Levels of MIDs Power Seizers (nonhereditary) 0.314 a 0.315 b 0.353 c 0.359 d (0.112) (0.089) (0.113) (0.091) Similar Constraints (nonhereditary) -0.081 a -0.015 b -0.049 c 0.015 d (0.103) (0.089) (0.104) (0.091) Power Seizers (hereditary) 0.364 e -0.047 f 0.417 g 0.039 h (0.175) (0.205) (0.179) (0.209) Similar Constraints (hereditary) 0.076 e 0.153 f 0.116 g 0.189 h (0.089) (0.076) (0.090) (0.078) Trade Openness -0.351 -0.088 -0.360 -0.038 -0.357 -0.096 -0.366 -0.047 (0.142) (0.087) (0.151) (0.099) (0.151) (0.088) (0.157) (0.101) Years since Regime Est. 0.001 0.004 t 0.001 0.004 t -0.001 0.004 t -0.001 0.005 t (0.003) (0.002) (0.003) (0.002) (0.003) (0.002) (0.003) (0.002) Capability 2.347 -0.143 2.316 -0.075 1.404 -0.710 1.332 -0.676 (1.152) (1.279) (1.159) (1.310) (1.185) (1.268) (1.211) (1.299) Total Borders 0.070 0.069 0.079 0.078 (0.017) (0.017) (0.017) (0.018) Major Power Status 0.395 13.692 0.406 20.618 0.436 13.736 0.447 13.439 (0.137) (508.492) (0.137) (.) (0.144) (628.335) (0.145) (474.635) Total Allies 0.012 -0.000 0.014 0.006 0.004 -0.002 0.007 0.004 (0.017) (0.016) (0.017) (0.016) (0.019) (0.016) (0.019) (0.016) Years Since Last MID -0.023 0.042 -0.022 0.043 -0.024 0.043 -0.023 0.043 (0.010) (0.013) (0.010) (0.013) (0.010) (0.013) (0.010) (0.013) Cubic Splines YES YES YES YES YES YES YES YES Region Dummies YES NO YES NO YES NO YES NO Constant -0.895 -0.868 -0.817 -0.791 (0.341) (0.344) (0.357) (0.361) Observations 7540 6922 7540 6874 7540 6922 7540 6874 Negative binomial regressions with standard errors in parentheses clustered by country. t p< 0.10, p< 0.05, p< .01, p< .001 Models (2), (4), (6), and (8) are fixed effects models of (1), (3), (5), and (7), respectively, and thus drop time-invariant variables such as contiguity (Total Borders) and region dummies. a Wald Test: different at p< 0.0004. b Wald Test: different at p< 0.0001. c Wald Test: different at p< 0.0004. d Wald Test: different at p< 0.0000. e Wald Test: different at p< 0.1070. f Wald Test: different at p< 0.3168. g Wald Test: different at p< 0.0947. h Wald Test: different at p< 0.4609. 137 the ones on the second column fall short of demonstrating statistical significance due to larger confidence intervals around the estimates. The major take-away point is that despite similar institutional constraints, dif- ferent leaders have different effects on a country’s decision to initiate a dispute. In other words, Iraqi invasion of Kuwait in 1990 can be attributed as much to the fact that Sad- dam Hussein was a Power Seizer as to the fact that he was an unconstrained strongman as suggested by the existing literature (Weeks 2012). Figure 9 reveals another interesting point of distinction between the constitu- tional and nonconstitutional Power Successors (row 3) - while the point estimates of the Power Seizers are very similar across the “nonhereditary” and the “hereditary” versions, those of the Power Seizers are much bigger for the "nonconstitutional" compared to the "constitutional". This suggests a pattern where the "nonconstitutional" Power Succes- sors are more aggressive than their counterparts in the "constitutional" dynasties. Recall that the "nonconstitutional" Power Successors usually do not go beyond one generation of power succession in most cases, or two at the most as are the cases of Nicaragua and North Korea (see Table 2). It implies that the immediate sons of the first generation dic- tators (Power Seizer (hereditary)) may be more peaceful than their fathers, but they are still more willing to engage in conflict compared to their counterparts in constitutional dynasties whose familial reign has been in place for more than decades and whose con- flict propensity has been averaged over a number of generations and over longer periods of time. Distinguishing Power Seizers and Comparing with Democracies Finally, how do these different leaders compare to democratic leaders? I con- duct an analysis by using the same leadership-type dummy variables as in the earlier 138 Figure 9: Expected Percentage of the Time That a Country Will Initiate Conflict: Iraq 1990 Scenario (Using CLARIFY) 139 analyses with control variables set to the values for Iraq in 1990, and set the base leadership-type category as “Democracy.” 112 I expect that the Power Seizers are more conflict prone than democratic leaders while the Power Successors are no more violent than democracies. The point estimates in Figure 10 show consistency with my expecta- tion about the Power Seizers, but the estimates differ for the Power Successors. Note that the Power Successors (constitutional) is the most peaceful by a large margin, even more so than democracies. However, the point estimate for the Power Successors (noncon- stitutional) suggests higher conflict propensity than democracies, and are statistically indistinguishable from the Power Seizers. This is consistent with the previous finding from Figure 9 that the immediate descendants from the Power Seizers tend to act still similar to their first generation ancestors in terms of conflict initiation compared to the remote descendants averaged over a longer time (i.e. constitutional Power Successors). In other words, although Kim Jung-un of North Korea may be less likely to initiate dis- pute compared to Kim Il-sung, he might be more willing to engage in conflict compared to his counterparts in other long established hereditary regimes. 112 Figure 10 is based off of model (1) of Table 13. 140 Figure 10: Expected Percentage of the Time That a Country Will Initiate Conflict: Iraq 1990 Scenario 141 Table 13: Power Seizers vs. Power Successors vs. Democracies (1) (2) FE (3) (4) FE Any MIDs Any MIDs Higher Levels of MIDs Higher Levels of MIDs Power Seizers(nonhereditary) 0.370 0.315 0.397 0.350 (0.114) (0.103) (0.115) (0.106) Power Seizers(hereditary) 0.406 -0.305 0.448 -0.248 (0.195) (0.231) (0.198) (0.235) Power Successors(constitutional) -0.606 -1.002 -0.569 -0.933 (0.172) (0.246) (0.182) (0.251) Power Successors(nonconstitutional) 0.112 -0.532 t 0.041 -0.608 t (0.310) (0.303) (0.250) (0.318) Other Autocrats 0.112 0.141 0.113 0.148 (0.108) (0.099) (0.105) (0.102) Trade Openness -0.346 -0.110 -0.357 -0.061 (0.151) (0.093) (0.161) (0.104) Years since Regime Est. 0.003 0.005 0.003 0.005 t (0.003) (0.002) (0.003) (0.002) Capability 2.390 -0.395 2.391 -0.276 (1.168) (1.272) (1.175) (1.302) Total Borders 0.066 0.065 (0.016) (0.016) Major Power Status 0.431 14.525 0.441 17.911 (0.141) (399.262) (0.138) (.) Total Allies 0.008 0.007 0.011 0.014 (0.017) (0.016) (0.017) (0.016) Years Since Last MID -0.022 0.042 -0.022 0.042 (0.010) (0.013) (0.010) (0.013) Cubic Splines YES YES YES YES Region Dummies YES NO YES NO Constant -0.955 -0.929 (0.342) (0.346) Observations 7490 6884 7490 6836 Negative binomial regressions with standard errors in parentheses clustered by country. t p< 0.10, p< 0.05, p< .01, p< .001 Models (2) and (4) are fixed effects of models of (1) and (3), respectively, and thus drop time-invariant variables such as contiguity (Total Borders) and region dummies. 142 Appendix Table 14: List of Power Seizers Table 14: List of Power Seizers Country Leader Start Date End Date Afghanistan Sardar Mohammad Daud Khan 17/07/1973 27/04/1978 Afghanistan Taraki 30/04/1978 27/03/1979 Afghanistan Karmal 27/12/1979 4/5/1986 Afghanistan Burhanuddin Rabbani 28/06/1992 27/09/1996 Afghanistan Mullah Omar 27/09/1996 13/11/2001 Albania Hoxha 29/11/1944 11/4/1985 Algeria Boumedienne 19/06/1965 27/12/1978 Algeria Kafi 2/7/1992 31/01/1994 Angola Neto 11/11/1975 10/9/1979 Argentina Peron 4/6/1946 20/09/1955 Argentina Aramburu 13/11/1955 1/5/1958 Argentina Ongania 28/06/1966 8/6/1970 Argentina Levingston 18/06/1970 22/03/1971 Argentina Lanusse 25/03/1971 25/05/1973 Argentina Videla 29/03/1976 29/03/1981 Azerbaijan Abulfaz Elchibey 16/06/1992 18/06/1993 Azerbaijan H. Aliyev 24/06/1993 4/8/2003 Bangladesh Sayem 6/11/1975 21/04/1977 Bangladesh Ziaur Rahman 21/04/1977 20/05/1981 Bangladesh Ershad 27/03/1982 6/12/1990 Bangladesh Ahmed 6/12/1990 20/03/1991 Benin Soglo 28/10/1963 19/01/1964 Benin Apithy 19/01/1964 29/11/1965 Benin Alley 22/12/1967 1/8/1968 Benin Zinsou 1/8/1968 10/12/1969 Benin Paul-Emile de Souza 13/12/1969 7/5/1970 Benin Kerekou 27/10/1972 4/4/1991 Bolivia Gutierrez 16/08/1946 10/3/1947 Bolivia Herzog 11/3/1947 7/5/1949 Bolivia Ballivian Rojas 16/05/1951 11/4/1952 Bolivia Paz Estenssoro 17/04/1952 17/06/1956 Bolivia Siles Zuazo 17/06/1956 6/8/1960 Bolivia Barrientos Ortuna 6/11/1964 5/1/1966 Bolivia Ovando Candia 26/09/1969 6/10/1970 Bolivia Torres 7/10/1970 22/08/1971 Bolivia Banzer Suarez 22/08/1971 21/07/1978 Bolivia Padilla Aranciba 24/11/1978 8/8/1979 143 Bolivia Garcia Meza Tejada 18/07/1980 4/8/1981 Bolivia Torrelio Villa 4/9/1981 19/07/1982 Brazil Vargas 10/2/1951 24/08/1954 Brazil Castello Branco 15/04/1964 15/03/1967 Brazil Costa de Silva 15/03/1967 1/9/1969 Bulgaria Dimitrov 23/11/1946 2/7/1949 Bulgaria Kolarov 20/07/1949 23/01/1950 Bulgaria Chervenkov 1/2/1950 16/04/1956 Burkina Faso Lamizana 3/1/1966 13/01/1971 Burkina Faso Zerbo 25/11/1980 7/11/1982 Burkina Faso J. P. Ouedraogo 7/11/1982 4/8/1983 Burkina Faso Sankara 4/8/1983 15/10/1987 Burkina Faso Campaore 15/10/1987 31/12/2004 Burundi Micombero 28/11/1966 1/11/1976 Burundi Bagaza 9/11/1976 3/9/1987 Burundi Buyoya 3/9/1987 10/7/1993 Cambodia Lon Nol 18/03/1970 10/4/1975 Cambodia Pol Pot 11/4/1975 7/1/1979 Cambodia Heng Samrin 7/1/1979 14/01/1985 Cambodia Hun Sen 14/01/1985 21/09/1993 Central African Republic Bokassa 1/1/1966 21/09/1979 Central African Republic Dacko 21/09/1979 1/9/1981 Central African Republic Kolingba 1/9/1981 22/10/1993 Central African Republic Patasse 22/10/1993 15/03/2003 Central African Republic Francois Bozize 30/04/2003 31/12/2004 Chad Malloum 15/04/1975 23/03/1979 Chad Oueddei 23/03/1979 7/6/1982 Chad Habre 19/06/1982 2/12/1990 Chad Deby 4/12/1990 31/12/2004 Chile Pinochet 11/9/1973 11/3/1990 China Chiang Kai-shek 26/12/1946 21/01/1949 China Mao Tse-Tung 1/10/1949 9/9/1976 Colombia Rojas Pinillia 13/06/1953 10/5/1957 Colombia Paris 10/5/1957 7/8/1958 Comoros Soilih 3/8/1975 13/05/1978 Comoros Abdallah 21/05/1978 26/11/1989 Comoros Djohar 26/11/1989 29/09/1995 Comoros Azali Assoumani 30/04/1999 21/01/2002 Congo Debat 16/08/1963 4/9/1968 Congo Raoul 4/9/1968 1/1/1969 Congo Ngouabi 1/1/1969 18/03/1977 Congo Nguesso 8/2/1979 20/08/1992 144 Congo Mobutu 25/11/1965 16/05/1997 Congo Laurent Kabila 16/05/1997 16/01/2001 Costa Rica Leon Herrera 20/04/1948 8/5/1949 Costa Rica Figueres Ferrer 8/11/1953 8/5/1958 Costa Rica Orlich 8/5/1962 8/5/1966 Costa Rica Quiros, Daniel 8/5/1974 8/5/1978 Cote d’Ivoire Laurent Gbagbo 26/10/2000 31/12/2004 Cuba Grau San Martin 10/10/1944 10/10/1948 Cuba Prio Socarres 10/10/1948 10/3/1952 Cuba Batista 10/3/1952 1/1/1959 Cuba Castro 2/1/1959 31/12/2004 Czechoslovakia Gottwald 7/6/1948 14/03/1953 Dominican Republic Rafel Trujillo 16/08/1930 30/05/1961 Dominican Republic Cabral 22/12/1963 25/04/1965 Ecuador Arosemena Monroy 7/11/1961 11/7/1963 Ecuador Castro 11/7/1963 29/03/1966 Ecuador Rodriguez Lara 15/02/1972 11/1/1976 Ecuador Poveda Burbano 11/1/1976 10/8/1979 Ecuador Gustavo Noboa 22/01/2000 15/01/2003 Ecuador Lucio Gutierrez 15/01/2003 31/12/2004 Egypt Naguib 23/07/1952 14/11/1954 Egypt Nasser 17/11/1954 28/09/1970 Egypt Sadat 28/09/1970 6/10/1981 El Salvador Cordova 15/12/1948 4/1/1949 El Salvador Oscar Bolanos 22/10/1949 14/09/1950 El Salvador Osorio 14/09/1950 14/09/1956 El Salvador Castillo 26/10/1960 25/01/1961 El Salvador Portillo 25/01/1961 25/01/1962 El Salvador Majano Ramos 15/10/1979 7/12/1980 El Salvador Duarte 13/12/1980 2/5/1982 El Salvador Magana Borjo 2/5/1982 1/6/1984 Equatorial Guinea Nguema Mbasogo 3/8/1979 31/12/2004 Eritrea Afeworki 24/05/1993 31/12/2004 Ethiopia Banti 28/11/1974 3/2/1977 Ethiopia Mengistu Marriam 11/2/1977 21/05/1991 Ethiopia Meles Zenawi 27/05/1991 31/12/2004 Fiji Rabuka 14/05/1987 19/05/1999 Gambia Jammeh 22/07/1994 31/12/2004 Georgia Shevardnadze 10/3/1992 23/11/2003 Georgia Burdjanadze 23/11/2003 25/01/2004 German Democratic Republic Wilhelm Pieck 22/04/1946 25/07/1950 Ghana Ankrah 24/02/1966 3/4/1969 145 Ghana Acheampong 13/01/1972 5/7/1978 Ghana Akuffo 5/7/1978 4/6/1979 Ghana Rawlings 31/12/1981 7/1/2001 Greece Sophoulis 8/9/1947 24/06/1949 Greece Papadopoulos 13/12/1967 25/11/1973 Greece Gizikis 25/11/1973 24/07/1974 Guatemala Arbenz Guzman 1/3/1951 27/06/1954 Guatemala Castillo Armas 8/7/1954 20/07/1957 Guatemala Flores Avendano 28/10/1957 2/3/1958 Guatemala Peralta Azurdia 31/03/1963 1/7/1966 Guatemala Arana Osorio 1/7/1970 1/7/1974 Guatemala Rios Montt 23/03/1982 8/8/1983 Guatemala Mejia Victores 8/8/1983 14/01/1986 Guinea Conte 3/4/1984 31/12/2004 Guinea-Bissau Cabral 10/9/1974 14/11/1980 Guinea-Bissau Vieira 14/11/1980 7/5/1999 Guinea-Bissau Sanha 14/05/1999 17/02/2000 Haiti Magloire 6/12/1950 12/12/1956 Haiti Pierre-Louis 12/12/1956 3/2/1957 Haiti Duvalier, Francois 15/10/1957 22/04/1971 Haiti Namphy 7/2/1986 7/2/1988 Haiti Avril 17/09/1988 10/3/1990 Haiti Cedras 30/09/1991 14/10/1994 Honduras Carias Andino 1/2/1933 1/1/1949 Honduras Hector Caraccioli 21/10/1956 21/12/1957 Honduras Lopez Arellano 3/10/1963 6/6/1971 Honduras Castro 22/04/1975 7/8/1978 Honduras Paz Garcia 7/8/1978 27/01/1982 Indonesia Suharto 12/3/1966 21/05/1998 Iran, Islamic Republic of Ayatollah Khomeini 1/2/1979 3/6/1989 Iran, Islamic Republic of Rafsanjani 17/08/1989 3/8/1997 Iraq Karrim Kassem 14/07/1958 8/2/1963 Iraq Salem Aref 8/2/1963 13/04/1966 Iraq Hassan Al-Bakr 17/07/1968 16/07/1979 Iraq Saddam Hussein 16/07/1979 9/4/2003 Iraq Paul Bremer 12/5/2003 28/06/2004 Japan Douglas MacArthur 2/9/1945 11/4/1951 Japan Matthew Bunker Ridgway 11/4/1951 28/04/1952 Korea, North Kim Il-Sung 9/9/1948 8/7/1994 Korea, Republic of Hee Park 3/7/1961 26/10/1979 Korea, Republic of Chun Doo Hwan 27/08/1980 25/02/1988 Korea, Republic of Roh Tae Woo 25/02/1988 25/02/1993 146 Lao Nosavan 31/12/1959 7/1/1960 Lao Boun Oum 13/12/1960 22/06/1962 Lao Phomivan 2/12/1975 25/11/1992 Lao Phounsavanh 25/11/1992 24/02/1998 Lao Siphandon 24/02/1998 31/12/2004 Lesotho Lekhanya 24/01/1986 2/5/1991 Lesotho Ramaema 2/5/1991 2/4/1993 Lesotho Mokhehle 2/4/1993 17/08/1994 Liberia Doe 12/4/1980 9/9/1990 Liberia Taylor 2/8/1997 11/8/2003 Libya Qaddafi 1/9/1969 31/12/2004 Macedonia Nikola Kljusev 27/01/1991 17/08/1992 Madagascar Ratsiraka 15/06/1975 27/03/1993 Madagascar Zafy 27/03/1993 5/9/1996 Madagascar Marc Ravalomanana 6/7/2002 31/12/2004 Malawi Banda 6/7/1964 21/05/1994 Mali Traore 19/11/1968 26/03/1991 Mali Amadou Toure 26/03/1991 6/6/1992 Mauritania Ould Mohamed Salek 10/7/1978 6/4/1979 Mauritania Ould Ahmed Louly 3/6/1979 4/1/1980 Mauritania Ould Haidalla 4/1/1980 12/12/1984 Mauritania Sidi Ahmed Taya 12/12/1984 31/12/2004 Mongolia Choibalsan 22/03/1936 26/01/1952 Mozambique Machel 25/06/1975 19/10/1986 Myanmar Ne Win 2/3/1962 25/07/1988 Myanmar Saw Maung 18/09/1988 23/04/1992 Myanmar Than Shwe 23/04/1992 31/12/2004 Nicaragua Anastasio Somoza Garcia 1/1/1937 1/5/1947 Nicaragua Daniel Ortega 18/07/1979 25/04/1990 Niger Kountche 17/04/1974 10/11/1987 Niger Mainassara 27/01/1996 11/4/1999 Niger Mamadou 22/12/1999 31/12/2004 Nigeria Gowon 29/07/1966 29/07/1975 Nigeria Ramat Mohammed 29/07/1975 13/02/1976 Nigeria Obasanjo 13/02/1976 1/10/1979 Nigeria Buhari 31/12/1983 27/08/1985 Nigeria Babangida 27/08/1985 26/08/1993 Nigeria Abacha 17/11/1993 8/6/1998 Pakistan Ayub Khan 7/10/1958 25/03/1969 Pakistan Zia 5/7/1977 17/08/1988 Pakistan Khalid 5/11/1996 17/02/1997 Pakistan Musharraf 14/10/1999 31/12/2004 147 Panama Arias, A. 25/11/1949 10/5/1951 Panama Remon Cantero 1/10/1952 2/1/1955 Panama Torrijos Herrera 12/10/1968 31/07/1981 Panama Florez Aguilar 31/07/1981 3/3/1982 Panama Dario Paredes 3/3/1982 15/08/1983 Panama Noriega 15/08/1983 3/1/1990 Paraguay Natalicio 15/08/1948 30/01/1949 Paraguay Chavez 10/9/1949 6/5/1954 Paraguay Stroessner 11/7/1954 3/2/1989 Paraguay Rodriguez Pedotti 3/2/1989 15/08/1993 Peru Bustamante 28/07/1945 28/10/1948 Peru Odria 29/10/1948 1/6/1950 Peru Perez Godoy 19/07/1962 3/3/1963 Peru Velasco Alvarado 3/10/1968 29/08/1975 Peru Morales Bermudez 30/08/1975 28/07/1980 Republic of Vietnam Ngo Dinh Diem 26/06/1954 1/11/1963 Republic of Vietnam Minh 1/11/1963 30/01/1964 Republic of Vietnam Van Huong 4/11/1964 26/01/1965 Republic of Vietnam Van Thieu 14/06/1965 21/04/1975 Romania Michael 23/08/1944 30/12/1947 Romania Roman 26/12/1989 1/10/1991 Russian Federation Stalin 10/3/1923 5/3/1953 Rwanda Kayibanda 26/10/1961 5/7/1973 Rwanda Habyarimana 5/7/1973 6/4/1994 Rwanda Paul Kagame 19/07/1994 31/12/2004 Sierra Leone Juxon-Smith 27/03/1967 19/04/1968 Sierra Leone Strasser 1/5/1992 17/01/1996 Sierra Leone Koroma 25/05/1997 12/2/1998 Singapore Lee Kuan Yew 5/6/1959 28/11/1990 Singapore Goh Chok Tong 28/11/1990 12/8/2004 Somalia Siad Barre 21/10/1969 26/01/1991 South Africa Mandela 10/5/1994 16/06/1999 Spain Franco 1/4/1939 30/10/1975 Sudan Abboud 18/11/1958 1/11/1964 Sudan al-Khalifa 1/11/1964 14/06/1965 Sudan Nimeiri 25/05/1969 19/07/1971 Sudan Abdul Rahman Swaredahab 6/4/1985 6/5/1986 Sudan Al-Bashir 30/06/1989 31/12/2004 Syrian Arab Republic Al-Hafiz 27/07/1963 25/02/1966 Syrian Arab Republic El-Atassi, N. 25/02/1966 13/11/1970 Syrian Arab Republic Al-Khatib 18/11/1970 22/02/1971 Syrian Arab Republic Al-Assad H. 22/02/1971 10/6/2000 148 Taiwan, Province of China Li Tsung-jen 21/01/1949 28/02/1950 Taiwan, Province of China Chiang Kai-shek 1/3/1950 5/4/1975 Tajikistan Nabiyev 23/09/1991 7/9/1992 Thailand Plaek Pibulsongkram 21/08/1946 16/09/1957 Thailand Thanon Kittakachorn 27/12/1957 20/10/1958 Thailand Sarit 20/10/1958 8/12/1963 Thailand Sanya 14/10/1973 21/02/1975 Thailand Sangad 6/10/1976 12/11/1977 Thailand Kriangsak 12/11/1977 29/02/1980 Togo Eyadema 14/04/1967 31/12/2004 Tunisia Ben Ali Bourguiba 25/07/1957 7/11/1987 Tunisia Zine Al-Abidine Ben Ali 7/11/1987 31/12/2004 Turkey Gursel 27/05/1960 10/11/1961 Turkey Erim 19/03/1971 17/04/1972 Turkey Evren 20/09/1980 23/11/1983 Uganda Amin 25/01/1971 11/4/1979 Uganda Okello 29/07/1985 29/01/1986 Uganda Museveni 29/01/1986 31/12/2004 Uruguay Mendez Manfredini 1/9/1976 1/9/1981 Uruguay Alvarez Armalino 1/9/1981 1/3/1985 Venezuela Betancourt 18/10/1945 15/02/1948 Venezuela Delgado Chalbaud 24/11/1948 13/11/1950 Venezuela Perez Jimenez 13/11/1950 23/01/1958 Venezuela Sanabria 14/11/1958 13/02/1959 Venezuela Hugo Chavez 2/2/1999 31/12/2004 Viet Nam Ho Chi Minh 2/9/1945 3/9/1969 Yemen Arab Republic AL-Sallal 27/09/1962 5/11/1967 Yemen Arab Republic Al-Hamadi 13/06/1974 11/10/1977 Yemen Arab Republic Al-Ghashmi 11/10/1977 24/06/1978 Yemen People’s Republic Al-Shaabi 30/11/1967 22/06/1969 Yemen People’s Republic Ali Rubayyi 24/06/1969 26/06/1978 Yemen People’s Republic Ismail 27/12/1978 2/4/1980 Zimbabwe Mugabe 4/3/1980 31/12/2004 —— 149 Appendix Figure 11: World Map by Authoritarian Leader Figure 11: World Map by Authoritarian Leader Label Definitions: FG(broad):Power Seizers(nonhereditary); FG(narrow): Power Seizers(hereditary); HS(broad): Power Successors(constitutional); HS(narrow): Power Successors(nonconstitutional) 150 Appendix Table 15: List of Power Successors (Constitu- tional) Table 15: List of Power Successors (Constitutional Country Leader Start Date End Date Afghanistan Mahmud Khan Ghazi 14/05/1946 7/9/1953 Afghanistan Sardar Mohammad Daud Khan 7/9/1953 10/3/1963 Afghanistan Zahir Shah 10/3/1963 17/07/1973 Bahrain Isa Ibn Al-Khalifah 15/08/1971 6/3/1999 Bahrain Hamad Isa Ibn Al-Khalifah 6/3/1999 -present Bhutan Wangchuk, Jigme 21/08/1926 24/03/1952 Bhutan Wangchuk, Jigme Dorji 24/03/1952 21/07/1972 Bhutan Wangchuck, Jigme Singye 21/07/1972 20/07/1998 Bhutan Lyonpo Jigme Thinley 20/07/1998 9/7/1999 Bhutan Lyonpo Sangay Ngedup 9/7/1999 20/07/2000 Bhutan Lyonpo Yeshey Zimba 20/07/2000 8/8/2001 Bhutan Lyonpo Khandu Wangchuk 8/8/2001 14/08/2002 Bhutan Lyonpo Kinzang Dorji 14/08/2002 30/08/2003 Bhutan Lyonpo Jigme Thinley 30/08/2003 18/08/2004 Bhutan Lyonpo Yeshey Zimba 18/08/2004 05/09/2005 Bhutan Lyonpo Sangay Ngedup 05/09/2005 07/09/2006 Bhutan Lyonpo Khandu Wangchuk 07/09/2006 31/07/2007 Burundi Mwambutsa 1/7/1962 8/7/1966 Egypt Farouk 28/04/1936 22/07/1952 Iran, Islamic Rep. Mohammad Reza 16/09/1941 30/04/1951 Iran, Islamic Rep. Mohammad Mossadeg 30/04/1951 19/08/1953 Iraq Abdul-Ilah 4/4/1939 2/5/1953 Iraq Faisal II 2/5/1953 14/7/1958 Jordan Abdullah Al-Hussein 11/4/1921 20/7/1951 Jordan Ibn Abdullah-Hashim 6/9/1951 11/8/1952 Jordan Hussein Ibn Talal El-Hashim 11/8/1952 7/2/1999 Jordan Abdullah Ibn Hussein El-Hashimi 7/2/1999 -present Kuwait Abdullah As-Sabah 29/01/1950 24/11/1965 Kuwait Sabah As-Sabah 24/11/1965 31/12/1977 Kuwait Jabir As-Sabah 1/1/1978 15/1/2006 Kuwait Sheikh Saad Al Abdullah Al Salim Al Sabah 15/1/2006 24/1/2006 Kuwait Sheikh Sabah Al Ahmad Al Jabir Al Sabah 1/24/2006 -present Libya Idris 24/12/1951 1/9/1969 Morocco Mohammed V 2/3/1956 26/02/1961 Morocco Hassan II 26/02/1961 23/07/1999 151 Morocco Muhammad VI 23/07/1999 -present Nepal Tribhuvan 18/02/1951 13/03/1955 Nepal Mahendra 14/03/1955 31/01/1972 Nepal Birendra 31/01/1972 1/6/2001 Nepal Lokendra Bahadur Chand 11/10/2002 5/6/2003 Nepal Surya Bahadur Thapa 5/6/2003 3/6/2004 Nepal Sher Bahdur Deuba 3/6/2004 1/2/2005 Oman Sa‘id ibn Taimur 10/2/1932 23/07/1970 Oman Qabus Bin Said 23/07/1970 -present Qatar Ahmed Ath-Thani 3/9/1971 22/02/1972 Qatar Khalifah Ath-Thani 22/02/1972 27/06/1995 Qatar Hamad ibn Khalifah Al Thani 27/06/1995 25/6/2013 Saudi Arabia Aziz 20/05/1927 9/11/1953 Saudi Arabia Saud 9/11/1953 2/11/1964 Saudi Arabia Faisal 2/11/1964 25/03/1975 Saudi Arabia Khalid 25/03/1975 13/06/1982 Saudi Arabia Fahd 13/06/1982 1/1/1996 Saudi Arabia Abdullah 1/1/1996 23/1/2015 Swaziland Subhuza II 6/9/1968 21/08/1982 Swaziland Dzeliwe Shongwe 21/08/1982 10/8/1983 Swaziland Ntombe Thwala 10/8/1983 25/04/1986 Swaziland Mswati 25/04/1986 -present UAE An-Nahayan 2/12/1971 31/12/2004 UAE Khalifa Al Nahayan 31/12/2004 -present Yemen Arab Rep. Ibn Yahya Hamid 13/03/1948 18/09/1962 152 Appendix Figure 12: Weeks’ 2012 Militarism and Per- sonalism Index Figure 12: Weeks’ Militarism and Personalism Index in the nonconstitutional Power Successor countries. As discussed in the main text, it is hard to isolate one effect from the other on autocrat’s conflict propensity. 153 Appendix Figure 13: GWF 2014 Classification of the nine nonconstitutional Power Successor countries Figure 13: GWF Classification of the nine nonconstitutional Power Successor coun- tries. The vertical red line marks the year when power succession took place. The nonconsti- tutional dynasties are classified as many different categories. 154 Appendix Table 16, 17, 18, 19: List of Disputes Under Chiang Kai-shek, Chiang Ching-kuo, Kim Il-Sung, Kim Jong-il I provide a list of disputes under Chiang Kai-Shek (1950-1974), Chiang Ching-kuo (1978-1987), Kim Il-Sung (1946-1993), and Kim Jong-il (1994-2010) based on Corre- lates of War Project, MIDv4.0 Project.(Maoz et al. 2018) Table 16: List of Disputes under Chiang Kai-Shek (1950-1974) Dispute No. Year Fatality Hostility Level Description (where available) 1374 1950 101-250 deaths Use of Force 2008 1950 None Use of Force 2992 1951 None Use of Force 2990 1952 None Use of Force 50 1953 26-100 deaths Use of Force 2001 1953 26-100 deaths Use of Force 2036 1953 None Use of Force 2988 1953 None Use of Force 3203 1953 None Use of Force 3204 1954 None Use of Force 1448 1954 None Use of Force 2987 1954 None Use of Force 2037 1955 None Use of Force 3211 1955 None Use of Force 3212 1955 None Use of Force 1452 1956 None Display of Force “Spratly Islands Dispute I” 2986 1956 None Use of Force 155 3221 1956 None Use of Force 2049 1957 Missing Use of Force 2935 1956 Missing Display of Force Table 17: Disputes Under Chiang Ching-kuo (1978-1987) Dispute No. Year Fatality Hostility Level Description (where available) 2782 1987 None Use of Force Table 18: Disputes Under Kim Il-Sung (1946-1993) Dispute No. Year Fatality Hostility Level Description (where available) 2186 1949 Missing Use of Force “Ongjin Peninsula” 51 1950 >999 deaths War “Korean War” 2244 1954 1-25 deaths Use of Force “Korean War Truce Violations” 2187 1958 1-25 deaths Use of Force 2188 1962 1-25 deaths Use of Force 2189 1963 1-25 deaths Use of Force 2916 1965 None Use of Force “Seizure of USS Pueblo” 347 1968 None Use of Force 2941 1969 None Use of Force 2190 1970 26-100 deaths Use of Force 2191 1974 1-25 deaths Use of Force 362 1976 1-25 deaths Use of Force “Tree Trimming Incident” 2192 1977 1-25 deaths Use of Force “North Korean Economic Zone” 156 2963 1977 None Use of Force 2193 1979 1-25 deaths Use of Force 2194 1981 None Use of Force 2974 1982 None Use of Force 2196 1984 None Use of Force 2198 1986 None Use of Force 3917 1989 None Use of Force 3570 1992 1-25 deaths Use of Force 4018 1993 26-100 deaths Use of Force “Between January and April 1993, North Korea provoked a series of 42 incidents with Chinese border guards. . . . ” 4022 1993 1-25 deaths Display of Force “. . . During the dispute, North Korea and the American-South Korean side exchanged alerts, threats of forces and war, and shows of force. Both sides conducted border fortifications and North Korea violated the South Korean border as it challenged the Korean War armistice agreement. “ 157 Table 19: Disputes Under Kim Jong-il(1994-2010) Dispute No. Year Fatality Hostility Level Description (where available) 4087 1994 26-100 deaths Use of Force “. . . In the dispute, North Korea continually challenges South Korean territorial boundaries along the DMZ and in the waters along the Northern Limit Line. South Korea responds by placing its forces on alert, conducting military exercises, showing ships and planes, and clashing with North Korean vessels that cross into South Korean waters and North Korea troops that cross the DMZ. Despite the distinction between the land and naval incursions by North Korea, their timing close together suggests a broader effort to challenge South Korean borders. . . . ” 4019 1995 None Use of Force “This dispute consisted of 1 incident in which a North Korean patrol boat fired at a Chinese fishing boat.” 158 4089 1997 Missing Use of Force “This dispute involved one incident between China and North Korea. In the dispute, North Korean and Chinese troops in the vicinity of Hyesan near the China-North Korea border clashed, exchanging gunfire.” 4226 1997 None Display of Force “This dispute consists of one incident in which North Korea deployed ballistic missiles on mobile launchers that were capable of hitting Tokyo. Reports indicated that the missile deployment was a North Korean attempt to improve their bargaining position in normalization talks with Japan.“ 159 4225 1999 None Display of Force “This dispute consists of three incidents in which North Korean spy vessels entered Japanese waters. In response, Japanese vessels chased the vessels for 24 hours. During the chase Japanese planes dropped bombs and Japanese naval vessels fired warning shots. In addition, Japan scrambled 2 F-15s to the area. In response, North Korea scrambled 4 fighters to the area. “ 4218 2000 None Threat to Use Force “This dispute consists of 2 incidents in which North Korea warned that if American or South Korean vessels go beyond designated lanes and waterways in the Yellow Sea it will be regarded as an intrusion into North Korean waters and would be subject to military retaliation. . . .” 160 4282 2001 1-25 deaths Use of Force “This dispute consists of 2 incidents in which North Korea warned that if American or South Korean vessels go beyond designated lanes and waterways in the Yellow Sea it will be regarded as an intrusion into North Korean waters and would be subject to military retaliation. . . .” 4473 2005 None Use of Force “Border guards in the Korean DMZ briefly exchanged fire. While PRK forces fired first, it is unclear why the exchange occurred.” 4479 2007 None Threat to Use Force “threatened to forcibly board a Chinese surveillance ship after the ship ignored warnings to leave territorial waters.” 4481 2008 None Display of Force “Continuing the dispute between North and South Korea over the NLL, North Korean ships briefly violated South Korean waters on several occasions, but these incursions resulted in no armed hostilities.” 161 4490 2010 None Use of Force “North Korean border patrol forces opened fire on a Chinese vessel suspected of smuggling in the rivers connecting North Korea and China. Three Chinese civilians were killed.“ 162 References Anderson, Lisa. 1991. “Absolutism and the Resilience of Monarchy in the Middle East.” Political Science Quarterly 106(1):1–15. Beck, Nathaniel, Jonathan N Katz and Richard Tucker. 1998. “Taking time seriously: Time-series-cross-section analysis with a binary dependent variable.” American Jour- nal of Political Science 42(4):1260–1288. Bellin, Eva. 2004. “The robustness of authoritarianism in the Middle East: Exception- alism in comparative perspective.” Comparative politics pp. 139–157. Bellin, Eva. 2012. “Reconsidering the robustness of authoritarianism in the Middle East: Lessons from the Arab Spring.” Comparative Politics 44(2):127–149. Bennett, Andrew. 2010. “Process tracing and causal inference.”. Bennett, Andrew and Alexander L George. 2005. “Case studies and theory development in the social sciences.” Cambridge, Mass . Bennett, Andrew and Colin Elman. 2006. “Complex causal relations and case study methods: The example of path dependence.” Political analysis 14(3):250–267. Bienen, Henry and Nicolas Van de Walle. 1991. Of time and power: Leadership dura- tion in the modern world. Stanford University Press. Boorman, Howard L, Joseph KH Cheng and Janet Krompart. 1967. Biographical dic- tionary of republican China. V ol. 3 Columbia University Press. Brady, Henry E, David Collier and Jason Seawright. 2004. Refocusing the discussion of methodology. In Rethinking social inquiry: Diverse tools, shared standards. Rowman and Littlefield. Brand, Laurie A. 2013. Jordan’s Inter-Arab Relations: The Political Economy of Alliance-Making. Columbia University Press. Brecher, Michael. 1996. “Crisis escalation: Model and findings.” International Political Science Review 17(2):215–230. Brownlee, Jason. 2007. “Hereditary succession in modern autocracies.” World Politics 59(04):595–628. Bueno de Mesquita, Bruce, Alastair Smith, RM Siverson and James D Morrow. 2003. “Logic of Political Survival Cambridge, Mass.”. Burling, Robbins. 1974. The passage of power. Academic Press. 163 Byman, Daniel L and Kenneth M Pollack. 2001. “Let us now praise great men: Bringing the statesman back in.” International Security 25(4):107–146. Campbell, Donald T et al. 6. “Stanley. JC 1966.” Experimental and quasi-experimental designs for research . Carter, David B and Curtis S Signorino. 2010. “Back to the future: Modeling time dependence in binary data.” Political Analysis 18(3):271–292. Chehabi, Houchang E and Juan J Linz. 1998. Sultanistic regimes. JHU Press. Cheibub, José Antonio, Jennifer Gandhi and James Raymond Vreeland. 2010. “Democ- racy and dictatorship revisited.” Public Choice 143(1-2):67–101. Ching-kuo, Chiang. 1989. Chiang Ching-kuo, Collected Works. V ol. 2 Taipei: Govern- ment Information Office. Cline, Ray S. 1989. Chiang Ching-kuo Remembered: The man and his political legacy. University Press of Amer. Colgan, Jeff. 2012. “Measuring revolution.” Conflict Management and Peace Science 29(4):444–467. Colgan, Jeff D. 2013a. “Domestic revolutionary leaders and international conflict.” World Politics 65(04):656–690. Colgan, Jeff D. 2013b. “Fueling the fire: Pathways from oil to war.” International Security 38(2):147–180. Colgan, Jeff D and Jessica LP Weeks. 2015. “Revolution, personalist dictatorships, and international conflict.” International Organization 69(01):163–194. Colgan, Jeffrey David. 2010. Oil, Revolution, and International Conflict: the origins of resource-backed aggression. Princeton University. Collier, David. 2011. “Understanding process tracing.” PS: Political Science & Politics 44(4):823–830. Collier, David, James Mahoney and Jason Seawright. 2004. “Claiming too much: Warn- ings about selection bias.” Rethinking social inquiry: Diverse tools, shared standards pp. 85–102. Crozier, Brian and Eric Chou. 1976. Man Who Lost China. SCRIBNER’S. Danilovic, Vesna and Joe Clare. 2004. Authoritarian Regimes and International Con- flict. In annual meeting of the International Studies Association Meeting, Montreal, Canada. 164 de Ridder, Denise TD, Gerty Lensvelt-Mulders, Catrin Finkenauer, F Marijn Stok and Roy F Baumeister. 2012. “Taking stock of self-control a meta-analysis of how trait self-control relates to a wide range of behaviors.” Personality and Social Psychology Review 16(1):76–99. Debs, Alexandre and Hein E Goemans. 2010. “Regime type, the fate of leaders, and war.” American Political Science Review 104(03):430–445. do Kang, Myong. 1995a. Pyeongyang Dreams of Refuge. Seoul:Joongang Daily News. do Kang, Myong. 1995b. “Testimony.” JoongAng Ilbo . Ezrow, Natasha M and Erica Frantz. 2011. Dictators and dictatorships: Understanding authoritarian regimes and their leaders. Bloomsbury Publishing USA. Fearon, James D. 1994. “Domestic political audiences and the escalation of international disputes.” American Political Science Review 88(03):577–592. Frederick, Shane. 2006. “Valuing future life and future lives: A framework for under- standing discounting.” Journal of Economic Psychology 27(5):667–680. Gandhi, Jennifer and Adam Przeworski. 2006. “Cooperation, cooptation, and rebellion under dictatorships.” Economics & Politics 18(1):1–26. Gause, F Gregory. 1994. Oil monarchies: domestic and security challenges in the Arab Gulf States. Council on Foreign Relations. Geddes, Barbara. 1990. “How the cases you choose affect the answers you get: Selection bias in comparative politics.” Political analysis 2:131–150. Geddes, Barbara. 1999a. Authoritarian breakdown: Empirical test of a game theoretic argument. In annual meeting of the American Political Science Association, Atlanta. V ol. 2. Geddes, Barbara. 1999b. “What do we know about democratization after twenty years?” Annual review of political science 2(1):115–144. Geddes, Barbara. 2003. Paradigms and sand castles: Theory building and research design in comparative politics. University of Michigan Press. Geddes, Barbara, Joseph Wright and Erica Frantz. 2014. “Autocratic breakdown and regime transitions: A new data set.” Perspectives on Politics 12(02):313–331. Gerring, John. 2001. Social science methodology: A criterial framework. Cambridge University Press. Gerring, John. 2007. “Is there a (viable) crucial-case method?” Comparative Political Studies 40(3):231–253. 165 Ginges, Jeremy and Scott Atran. 2011. “War as a moral imperative (not just practical politics by other means).” Proceedings of the Royal Society of London B: Biological Sciences 278(1720):2930–2938. Glad, Betty. 2002. “Why tyrants go too far: Malignant narcissism and absolute power.” Political Psychology 23(1):1–2. Gleditsch, Kristian Skrede. 2002. “Expanded trade and GDP data.” Journal of Conflict Resolution 46(5):712–724. Gleditsch, Kristian Skrede, Idean Salehyan and Kenneth Schultz. 2008. “Fighting at home, fighting abroad: How civil wars lead to international disputes.” Journal of Conflict Resolution . Goertz, Gary. 2006. Social science concepts: A user’s guide. Princeton University Press. Gowa, Joanne. 2000. Ballots and bullets: The elusive democratic peace. Princeton University Press. Gurr, Ted Robert. 1988. “War, revolution, and the growth of the coercive state.” Com- parative Political Studies 21(1):45–65. Herb, Michael. 1999. All in the family: absolutism, revolution, and democracy in Middle Eastern monarchies. SUNY Press. Herz, John H. 1952. “The problem of successorship in dictatorial regimes; A study in comparative law and institutions.” The journal of Politics 14(01):19–40. Horowitz, Michael C and Allan C Stam. 2014. “How prior military experience influences the future militarized behavior of leaders.” International Organization 68(03):527–559. Huntington, Samuel P. 1968. “The bases of accommodation.” Foreign Affairs 46(4):642–656. Huth, Paul K. 1996. “Enduring rivalries and territorial disputes, 1950-1990.” Conflict Management and Peace Science 15(1):7–41. Hymans, Jacques EC. 2012. Achieving nuclear ambitions: scientists, politicians, and proliferation. Cambridge University Press. Jung, C.Y . 2014. An analysis of North Korean Power Succession Process: Kim Jong Il and Kim Jong Un cases PhD thesis Korea National Defense University. Kai-shek, Chiang. 1942. Chiang Diaries. V ol. box 42. 166 Kertzer, Joshua D. 2013. “Making sense of isolationism: foreign policy mood as a multilevel phenomenon.” The Journal of Politics 75(01):225–240. Kertzer, Joshua D, Kathleen E Powers, Brian C Rathbun and Ravi Iyer. 2014. “Moral support: How moral values shape foreign policy attitudes.” The Journal of Politics 76(03):825–840. Kim, Myong Chol. 1982. “Biography of an Infant Prodigy.” Far Eastern Economic Review . Kim, Sung Il. 1995. With the Century. Foreign Languages Publishing House. King, Gary, Robert O Keohane and Sidney Verba. 1994. Designing social inquiry: Scientific inference in qualitative research. Princeton university press. Kreps, Sarah. 2008. “When does the mission determine the coalition? The logic of multilateral intervention and the case of Afghanistan.” Security Studies 17(3):531– 567. Lai, Brian and Dan Slater. 2006. “Institutions of the offensive: Domestic sources of dispute initiation in authoritarian regimes, 1950–1992.” American Journal of Political Science 50(1):113–126. Lee, S.T. 2005. Kim Jong-il’s Successor: Through Kim Jong-il’s Succession Experience PhD thesis Dongkuk University. Leeds, Brett Ashley. 2003. “Do alliances deter aggression? The influence of military alliances on the initiation of militarized interstate disputes.” American Journal of Po- litical Science 47(3):427–439. Lerner, Daniel. 1958. “The passing of traditional society: Modernizing the Middle East.”. Levy, Jack S. 2008. “Case studies: Types, designs, and logics of inference.” Conflict management and peace science 25(1):1–18. Levy, Jack S and Lily I Vakili. 2014. “Diversionary action by authoritarian regimes: Ar- gentina in the Falklands/Malvinas case.” The internationalization of communal strife pp. 118–46. Lewis, Bernard. 2000. “Monarchy in the Middle East.” Middle East Monarchies. The Challenge of Modernity, Boulder: Lynne Rienner pp. 15–22. Lim, Jae-Cheon. 2012. “North Korea’s Hereditary Succession: Comparing Two Key Transitions in the DPRK.” Asian Survey 52(3):550–570. 167 Linz, Juan J and Alfred Stepan. 1996. Problems of democratic transition and consoli- dation: Southern Europe, South America, and post-communist Europe. JHU Press. Loh, Pichon Pei Yung. 1971. The Early Chiang Kai-shek: A Study of His Personality and Politics, 1887-1924. New York: Columbia University Press. Lucas, Russell E. 2004. “Monarchical authoritarianism: Survival and political liberal- ization in a Middle Eastern regime type.” International Journal of Middle East Studies 36(01):103–119. Luciani, Giacomo. 1990. “Allocation vs. production states: A theoretical framework.” The Arab State pp. 65–84. Mao. 1965. Mr. Chiang Kai-shek, 1887-1926. Hong Kong:Longman. Maoz, Zeev, Paul L Johnson, Jasper Kaplan, Fiona Ogunkoya and Aaron P Shreve. 2018. “The dyadic militarized interstate disputes (MIDs) dataset version 3.0: Logic, char- acteristics, and comparisons to alternative datasets.” Journal of Conflict Resolution p. 0022002718784158. Marshall, Monty G. 1999. Third world war. Rowman & Littlefield Publishers. Marshall, Monty G, Keith Jagger and Ted Robert Gurr. 2009. “Polity IV: Regime author- ity characteristics and transition datasets, 1800-2009 [Data file].” Available on-line at http://www.systemicpeace.org/inscrdata.html . Martin, Bradley K. 2004. Under the loving care of the fatherly leader: North Korea and the Kim dynasty. Macmillan. McDermott, Rose. 2001. Risk-taking in international politics: Prospect theory in Amer- ican foreign policy. University of Michigan Press. Menaldo, Victor. 2012. “The middle east and north Africa’s resilient monarchs.” The Journal of Politics 74(03):707–722. O, Tae-Chin. 1990. “Three Personalities Rising as North Korea’s Next Generation of Leaders.” Chugan Choson . Oneal, John R and Bruce Russett. 2001. “Clear and clean: The fixed effects of the liberal peace.” International Organization 55(02):469–485. Oneal, John R, Bruce Russett and Michael L Berbaum. 2003. “Causes of peace: Democ- racy, interdependence, and international organizations, 1885–1992.” International Studies Quarterly 47(3):371–393. Palmer, Glenn, Vito d’Orazio, Michael Kenwick and Matthew Lane. 2015. “The MID4 dataset, 2002–2010: Procedures, coding rules and description.” Conflict Management and Peace Science 32(2):222–242. 168 Peceny, Mark and Caroline C Beer. 2003. “Peaceful parties and puzzling personalists.” American Political Science Review 97(02):339–342. Peceny, Mark, Caroline C Beer and Shannon Sanchez-Terry. 2002. “Dictatorial peace?” American Political Science Review 96(01):15–26. Peceny, Mark and Christopher K Butler. 2004. “The Conflict Behavior of Authoritarian Regimes1.” International Politics 41(4):565–581. P’eng-fei, Ch’I. 1994. History of the Chiang Kai-shek Family. Peking: Tuan Chieh Press. Posen, Barry R. 1984. “The sources of military doctrine: France.” Britain, and Germany between the World Wars (Ithaca, NY: Cornell UP 1984) . Ragin, Charles C. 2000. Fuzzy-set social science. University of Chicago Press. Rathbun, Brian C, Joshua D Kertzer and Mark Paradis. 2014. “Homo Diplomaticus: Mixed-Method Evidence of Variation in Strategic Rationality.” Manuscript, Univer- sity of Southern California . Reiter, Dan and Allan C Stam. 2002. Democracies at war. Princeton University Press. Rosen, Stephen Peter. 2009. War and human nature. Princeton University Press. Sambanis, Nicholas. 2004. “What is civil war? Conceptual and empirical complexities of an operational definition.” Journal of conflict resolution 48(6):814–858. Saunders, Elizabeth Nathan. 2011. Leaders at war: how presidents shape military in- terventions. Cornell University Press. Schlumberger, Oliver. 2010. “Opening old bottles in search of new wine: On nondemo- cratic legitimacy in the Middle East.” Middle East Critique 19(3):233–250. Schultz, Kenneth A. 2001. Democracy and coercive diplomacy. V ol. 76 Cambridge University Press. Scott Long, J. 1997. “Regression models for categorical and limited dependent vari- ables.” Advanced quantitative techniques in the social sciences 7. Seawright, Jason and John Gerring. 2008. “Case selection techniques in case study re- search: A menu of qualitative and quantitative options.” Political Research Quarterly 61(2):294–308. Sechser, Todd S. 2004. “Are soldiers less war-prone than statesmen?” Journal of Con- flict Resolution 48(5):746–774. 169 Siverson, Randolph M, James D Morrow, Bruce Bueno de Mesquita and Alastair Smith. 2003. The Logic of Political Survival. MIT Press. Skocpol, Theda. 1979. States and social revolutions: A comparative analysis of France, Russia and China. Cambridge University Press. Snyder Glenn, H. 1997. “Alliance Politics.”. Snyder, Jack. 1984. “Civil-Military Relations and the Cult of the Offensive, 1914 and 1984.” International Security 9(1):108–146. SRLee. 2015. “An Analysis of Kim Jong-un Through His Succession Process.” National Assembly Research Service 259. Stilwell, General Joseph W. 1991. The Stilwell Papers. Da Capo Press. Suh, Dae-Sook. 1995. Kim Il Sung: The North Korean Leader. Columbia University Press. Tannenwald, Nina. 1999. “The nuclear taboo: The United States and the normative basis of nuclear non-use.” International organization 53(3):433–468. Taylor, Jay. 2009a. The Generalissimo. Harvard University Press. Taylor, Jay. 2009b. The Generalissimo’s son: Chiang Ching-kuo and the revolutions in China and Taiwan. Harvard University Press. Teorell, Jan and Axel Hadenius. 2007. “Pathways from authoritarianism.” Journal of democracy 18(1):143–157. Tetlock, Philip E. 2003. “Thinking the unthinkable: Sacred values and taboo cogni- tions.” Trends in cognitive sciences 7(7):320–324. Todd Spinks, B, Emile Sahliyeh and Brian Calfano. 2008. “The Status of Democracy and Human Rights in the Middle East: Does Regime Type Make a Difference?” Democratisation 15(2):321–341. Tomz, Michael, Jason Wittenberg, Gary King et al. 2003. “CLARIFY: Software for interpreting and presenting statistical results.” Journal of Statistical Software 8(1):1– 30. Tong, Hollington Kong. 1953. Chiang Kai-shek. China Pub. Company. Tullock, Gordon. 1987. Autocracy. Springer Science & Business Media. Wahman, Michael, Jan Teorell and Axel Hadenius. 2013. “Authoritarian regime types revisited: updated data in comparative perspective.” Contemporary Politics 19(1):19– 34. 170 Walter, Barbara F. 2004. “Does conflict beget conflict? Explaining recurring civil war.” Journal of Peace Research 41(3):371–388. Wang. 1995. President Chiang Chin-kuo’s Late Years: Hao Pei-ts’un Diary. Taipei: Tien Hsia Literary Press. Weeks, Jessica L. 2008. “Autocratic audience costs: Regime type and signaling resolve.” International Organization 62(01):35–64. Weeks, Jessica L. 2012. “Strongmen and straw men: Authoritarian regimes and the initiation of international conflict.” American Political Science Review 106(02):326– 347. Yom, Sean L. 2013. “Jordan: The ruse of reform.” Journal of Democracy 24(3):127– 139. Yom, Sean L and F Gregory Gause III. 2012. “Resilient royals: How Arab monarchies hang on.” Journal of Democracy 23(4):74–88. 171

Abstract (if available)

Abstract It is a widely held belief that unconstrained personalist leaders tend to be most belligerent in their foreign policy pursuit due to the domestic institutional failure of punishing leaders who dominate the military and the state bureaucratic apparatus. This dissertation shows that not the domestic institutional constraints on the leader, but a leader’s predisposition towards use of force better explains the initiation of international military conflicts by autocrats. Using a panel data from 1946-2010, I argue that hereditary personalist autocrats tend to be less conflict prone than other personalist leaders even under similar institutional constraints. The “Power Successors” whose power acquisition came through hereditary ascription are significantly less conflict-prone than the “Power Seizers” who came to power through use of force even under similar institutional conditions. Furthermore, this dissertation employs two case studies—Taiwan as a typical case and North Korea as a divergent case—and demonstrates that the divergent paths to power experienced by the “Power Seizers” and the “Power Successors” explain and shape leaders’ beliefs about use of force.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Perspectives on state capacity and the political geography of conflict

PDF

Anti-foreign boycotts as a tool of economic coercion: the case of China

PDF

Habits and policy: the socio-cognitive foundations of foreign policy stability

PDF

Public opinion and international affairs: a multi-method approach to foreign policy attitudes

PDF

Volatile nationalism: nationalism and its influence on maritime disputes

PDF

Politics is something we do together: identity and institutions in U.S. elections

PDF

Circuit breakers: how policy entrepreneurs interrupted the electric flow with Peru’s first renewable energy legislation for the grid

PDF

The political economy of implementation: intellectual property rights protection across the world

Asset Metadata

Creator Shin, Jihyun (author)

Core Title Like father, like son? A succession-based explanation for conflict initiation by authoritarian regimes

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Political Science and International Relations

Publication Date 04/28/2019

Defense Date 03/05/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag authoritarian leader,conflict initiation,hereditary regimes,OAI-PMH Harvest,political psychology

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Kang, David (committee chair), Rathbun, Brian (committee member), Walsh, David (committee member), Weller, Nicholas (committee member)

Creator Email jihyun.shin@gmail.com,jihyunsh@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-150458

Unique identifier UC11660831

Identifier etd-ShinJihyun-7304.pdf (filename),usctheses-c89-150458 (legacy record id)

Legacy Identifier etd-ShinJihyun-7304.pdf

Dmrecord 150458

Document Type Dissertation

Format application/pdf (imt)

Rights Shin, Jihyun

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA